How to Write Effective AI Image Prompts
A well-written prompt is the key to getting stunning, controllable results from AI image generators. This guide explains the structure, techniques, and best practices that help you turn an idea into a high-quality image.
The Anatomy of a Great Prompt
An effective image prompt usually contains the following building blocks. You don't have to use all of them every time, but the more specific you are, the more predictable the result.
1. Subject
Describe the main thing you want to see. Be specific about who or what it is and what they are doing.
- Weak: a cat
- Strong: a fluffy orange tabby cat sitting on a windowsill, looking out at a rainy street
2. Environment / Background
Set the scene. Where is the subject? What is the atmosphere?
- indoor, outdoor, forest, city street, outer space, underwater
- rainy, sunny, foggy, neon-lit, golden hour
3. Art Style & Medium
Tell the AI how you want the image to look.
- Styles: photorealistic, cinematic, anime, oil painting, watercolor, pixel art, 3D render, minimalism
- Mediums: digital art, photograph, illustration, sketch, sculpture
4. Lighting
Lighting changes the mood dramatically.
- soft natural light, dramatic side lighting, backlit silhouette, studio lighting, volumetric fog, neon glow
5. Composition & Camera
Guide the framing and perspective.
- close-up portrait, wide-angle shot, bird's-eye view, low angle, rule of thirds, macro
- camera lens hints: 35mm, 85mm portrait lens, fisheye, shallow depth of field
6. Color Palette
Describe the dominant colors or overall tone.
- warm pastel tones, high contrast black and white, teal and orange, muted earth tones, vibrant psychedelic colors
7. Quality Modifiers
Add keywords that push the model toward higher fidelity.
- highly detailed, 8k, masterpiece, best quality, sharp focus, intricate, professional
Prompt Formula
You can use this simple template as a starting point:
[Subject] + [Action/State] + [Environment] + [Style/Medium] + [Lighting] + [Composition] + [Quality modifiers]Example:
A young woman in a red coat walking through a snowy Tokyo street at night, neon signs reflecting on wet pavement, cinematic composition, soft bokeh, highly detailed, 8k, photorealistic
Before and After Examples
Example 1
- Weak prompt: a dragon
- Strong prompt: a massive emerald dragon perched on a crumbling castle tower, stormy sky with lightning, dark fantasy art style, dramatic lighting, highly detailed scales, epic composition
Example 2
- Weak prompt: food
- Strong prompt: a gourmet burger on a rustic wooden board, melted cheese dripping, fresh lettuce and tomato, soft natural window light, food photography, shallow depth of field, appetizing, 8k
Example 3
- Weak prompt: forest
- Strong prompt: a mystical bioluminescent forest at twilight, glowing mushrooms and fireflies, misty atmosphere, fantasy illustration, vibrant teal and purple colors, wide-angle view, highly detailed
Platform-Specific Tips
Midjourney
- Use comma-separated descriptive keywords.
- Add parameters like
--ar 16:9,--style raw,--stylize 250. - Keep the prompt concise but evocative.
Example:
futuristic cyberpunk street market, neon lights, crowd of androids, rain, cinematic lighting, highly detailed --ar 16:9 --style raw
DALL-E
- Use clear, natural language.
- Avoid negative prompts; describe what you want instead.
- Works well with art styles and specific scene descriptions.
Example:
A watercolor painting of a cozy cottage by a lake, surrounded by autumn trees, soft morning light, peaceful atmosphere
Stable Diffusion
- Include quality tags such as
masterpiece,best quality,highly detailed. - Negative prompts help a lot:
blurry, low quality, deformed hands, extra fingers. - Use weighting for emphasis:
(detailed:1.2).
Flux
- Works well with natural, photographic descriptions.
- Emphasize lighting, texture, and realism.
Example:
A professional portrait of an elderly craftsman in his workshop, warm tungsten lighting, wood shavings on the table, shallow depth of field, photorealistic, 8k
Common Mistakes to Avoid
- Too vague: "a beautiful scene" gives the AI too much freedom.
- Too long and contradictory: conflicting styles confuse the model.
- Ignoring the subject's relationship to the scene: describe how they interact with the environment.
- Forgetting quality modifiers: without them, images can look soft or generic.
- Using copyrighted names or characters: may be blocked or produce inconsistent results.
Use ImagePromptify to Get Better Prompts
If you have a rough idea but struggle to describe it, try our built-in tools:
- Magic Enhance: expands a simple prompt into a detailed, professional version.
- Translate: converts your prompt into English for the best results with most AI image models.
Start with a simple sentence, let the tools refine it, and then adjust the details to match your vision.
Quick Checklist
- Is the subject clearly described?
- Did I mention the environment or background?
- Did I specify the art style or medium?
- Did I include lighting or atmosphere?
- Did I add quality modifiers for sharper results?
- Did I avoid contradictory descriptions?
Keep practicing, and soon writing great prompts will feel as natural as describing an image to a friend.