How to Write Prompts for AI Coloring Pages (With 40 Copy-Paste Examples)

Short answer: a good coloring-page prompt names a specific, visual subject first, then the style (“black and white line art coloring page, clean closed outlines, white background”), then the simplicity level for the audience (“thick bold outlines, large simple shapes” for young kids; “intricate, detailed” for adults), and finally the composition (“centered, single subject” or “full-page scene”). Describe what you want rather than what you don’t — image models are bad at negatives — and describe characters instead of naming them. That structure works in ChatGPT/DALL·E, Midjourney, Stable Diffusion, our own ColorThese generator, and whatever comes next. The reasoning and 40 ready-to-use prompts follow.

The four-part structure

1. Subject, concretely. “A fox” is weak. “A sleepy fox curled up under a crescent moon with a few stars” gives the model something to draw. One subject or one scene per prompt.
2. Style words. This is what turns a picture into a coloring page: black and white line art, coloring page, clean closed outlines, white background. Two or three of these phrases are plenty; stacking ten doesn’t help.
3. Simplicity level. The single biggest lever for usefulness. Thick bold outlines, large simple shapes, no small details for toddlers; medium detail for kids; intricate, highly detailed, fine line work for adults.
4. Composition. Centered single subject with plenty of white space, or full-page scene that extends to the edges, or symmetrical mandala. Say it, or the model picks for you.

A complete prompt reads like: “A friendly sea turtle swimming over a coral reef with three small fish, black and white line art coloring page, clean closed outlines, white background, medium detail, full-page scene.” That’s all it takes.

The rules that matter more than people expect

Describe, don’t negate

Diffusion models weight the nouns in a prompt; “no shading” still puts “shading” in the model’s head, and “NO extra arms” has been known to produce extra arms. Say “pure black lines on white, flat, unshaded” rather than “no grey, no color.” One or two light negatives rarely hurt, but they are not where the control is — positive description is. If your tool has a separate negative prompt field (Stable Diffusion, some others), put “color, shading, grey, gradient, text, watermark” there instead of in the main prompt.

Describe characters instead of naming them

Two reasons. Named characters are trademarks and copyrighted designs, which matters the moment you share or sell the page. And most generators now refuse or blur them — you get a generic blob. Describe the visual instead: “a small round green dinosaur with big friendly eyes” or “a brave girl in a blue dress with a bow and arrow.” You’ll get something original that still reads as the idea.

Keep text out of the image

Models garble letters. If you want a title or a name on the page, add it afterwards in an editor. Ask the generator for the picture only, and say “no text” if the tool tends to add captions — this is one negative that earns its place.

Know what models still get wrong

Hands and fingers, exact counts (“five ducks” often yields four or six), overlapping crowds, and tiny faces. Design around it: one or two characters, hands hidden or holding something, and counts you don’t care about. Simpler subjects also print better and color more easily.

Change one thing at a time

When a result is close but wrong, edit one element — the subject’s pose, the detail level, the composition — and regenerate. Rewriting the whole prompt makes it impossible to learn what helped. If your tool lets you fix a seed, keep it fixed while iterating on wording.

Ask for the page size you’ll print

Portrait 3:4 or letter proportions for a standard page; square for a mandala; landscape for a scene across a spread. If the tool takes pixel sizes, around 2550 × 3300 px prints a clean letter page; anything under 1500 px on the short side will look soft on paper. Our printing guide covers the rest.

40 prompts you can paste

Each one follows the structure above. Swap the subject and keep the rest. The style tail — “black and white line art coloring page, clean closed outlines, white background” — is implied after every subject below; add it in tools that need it spelled out every time.

Toddlers and preschool (thick lines, one big subject)

1. One large happy sun with a simple smiling face, very thick bold outlines, large simple shapes, no small details, centered with lots of white space.
2. A single big apple with one leaf, very thick bold outlines, large simple shape, centered.
3. One friendly round fish with three big scales, thick bold outlines, simple, centered.
4. A big simple teddy bear sitting down, thick bold outlines, large shapes, no background.
5. One chunky toy car seen from the side, thick bold outlines, big wheels, simple, centered.
6. A single big balloon on a string, thick outlines, large simple shape, centered.
7. A big friendly duck standing, thick bold outlines, large simple shapes, no background.
8. One large cupcake with a cherry on top, thick bold outlines, simple shapes, centered.

Kids 5–9 (medium detail, clear scenes)

9. A friendly T-rex walking through a jungle with two palm trees and a volcano, medium detail, full-page scene.
10. A puppy playing with a ball in a garden with a fence and flowers, medium detail, full-page scene.
11. A rocket ship flying past the moon with stars and one small planet, medium detail, centered.
12. A unicorn standing on a hill with a rainbow and clouds, medium detail, full-page scene.
13. A race car on a track with checkered flags, medium detail, side view, full-page.
14. A sea turtle swimming over a coral reef with three small fish, medium detail, full-page scene.
15. A knight in armor holding a shield standing in front of a castle, medium detail, centered.
16. A cheerful robot waving, with buttons and antennae, medium detail, centered, simple background.

Adults (intricate, fine line work)

17. A symmetrical mandala with eight-fold symmetry of lotus petals and leaves, intricate, fine line work, square composition, centered.
18. An ornate owl perched on a branch with detailed feather patterns and paisley motifs, intricate, centered.
19. A secret garden gate overgrown with roses, ivy and climbing flowers, highly detailed, full-page scene.
20. A tranquil koi pond with lily pads, ripples and two koi fish, intricate, full-page scene.
21. A vintage bicycle with a basket of flowers leaning on a brick wall, detailed line work, centered.
22. A mountain lake at dawn with pine trees and a small cabin, detailed, full-page landscape.
23. A stained-glass style hummingbird among trumpet flowers, bold leading lines with fine interior detail, centered.
24. A botanical study of a sunflower with leaves and seed pattern, detailed, centered.

Seasonal and holidays

25. A cozy fireplace with stockings, a wreath and a sleeping cat, medium detail, full-page scene.
26. A carved pumpkin on a porch step with a small friendly ghost and autumn leaves, medium detail, centered.
27. A basket of decorated eggs in spring grass with tulips and a bunny, medium detail, centered.
28. A snowman with a scarf and hat, pine trees and falling snowflakes, simple, full-page scene.
29. A heart wreath made of small flowers and ribbons, medium detail, centered, symmetrical.
30. A beach scene with a sandcastle, a bucket, a starfish and gentle waves, medium detail, full-page.
31. A turkey with fanned tail feathers decorated with leaf patterns, medium detail, centered.
32. Fireworks bursting over a small town skyline, medium detail, full-page scene.

Personal and custom

33. A golden retriever with a bandana sitting in front of a small house with a red door, medium detail, centered. (Swap in your pet and your house.)
34. A girl with curly hair and glasses riding a bicycle past a park bench, medium detail, full-page scene. (Describe the child; don’t upload strangers’ photos.)
35. A birthday cake with seven candles surrounded by balloons and streamers, medium detail, centered. (Change the number; check it came out right.)
36. A camper van parked by a lake with a campfire and two camp chairs, medium detail, full-page scene.
37. A classroom with a chalkboard, a globe and a row of desks, simple, full-page scene.
38. A family of four cartoon penguins wearing scarves, simple, centered.
39. A tractor in a field with a barn and a row of sunflowers, medium detail, full-page scene.
40. A city skyline with a bridge and a river, simplified shapes, full-page, landscape.

Adapting prompts for specific tools

• ChatGPT / DALL·E: conversational phrasing works; you can also say “make it simpler” or “thicker lines” in a follow-up and it will revise. Ask for a “printable” page and specify portrait.
• Midjourney: add style words like --style raw and an aspect ratio such as --ar 3:4; “coloring book page, line art, white background” is enough, and keep the subject up front.
• Stable Diffusion and similar: put color/shading/text/watermark in the negative prompt; use a line-art-oriented model or LoRA if you have one.
• ColorThese: the generator already appends the line-art styling, so just write the subject, the simplicity level and the composition — “a friendly sea turtle over a coral reef, medium detail, full-page scene” — and it produces clean printable line art. It also accepts a photo if you’d rather start from one (see turning a photo into a coloring page).

Checking a result before you print

Squint at it: one clear subject? Lines solid black, background white? Any grey shading that will fight the crayons? Any hands, counts or letters that came out wrong? Is the detail level right for who’s coloring (our toddler page guide is the strict version of this test)? Then download the full-size file and print at 100%.

Frequently asked questions

What is the best prompt for coloring pages?

Subject + “black and white line art coloring page, clean closed outlines, white background” + a simplicity level + a composition. Specific subject, positive description, no named characters.

How do I make AI coloring pages with thicker lines?

Ask for “very thick bold outlines, large simple shapes” and, if the tool supports it, a lower detail level. If lines still come out thin, thicken them in an image editor (a “Minimum” or “dilate” filter) before printing.

Why does the AI add grey shading or color?

The style words aren’t strong enough or the model defaults to illustration. Lead with “black and white line art coloring page” and “pure black lines on white, flat”; use a negative-prompt field if there is one; and try a tool built for line art.

Can I use a character from a movie or game in my prompt?

You can type it, but most tools will refuse or distort it, and the result is someone else’s intellectual property either way. Describe the look you want instead and you’ll get something original.

Can I sell AI-generated coloring pages?

Sometimes, with caveats about originality, platform rules and copyright status of AI output that vary by country and platform — we cover that in a separate guide on selling coloring pages.

Start with one of the 40 above, change the subject, and keep the structure. The difference between a frustrating generator and a useful one is mostly in those four parts.