How to Write Better AI Image Prompts: Midjourney & Stable Diffusion
Getting the right image from AI takes more than just describing what you see. You need to guide the style, lighting, composition, and technical details. This guide shows you the exact framework to write image prompts that work—whether you're using Midjourney, DALL-E, or Stable Diffusion.
Last updated · By the Prompt Orange team
Common mistakes to avoid
Too vague on style
Specify the visual style explicitly: 'editorial photo', '3D render', 'watercolor illustration', 'flat design icon'
Forgetting aspect ratio
Always include aspect ratio at the end: --ar 16:9, --ar 4:5, --ar 1:1
No lighting direction
Describe lighting: 'soft diffused light', 'dramatic side lighting', 'golden hour', 'studio lighting'
Conflicting descriptors
Avoid contradictions like 'minimalist detailed scene' or 'bright dark mood'—pick one clear direction
Before & after: Real example
See exactly how to transform a weak prompt into a strong one
“a woman in a park”
Too vague—AI has to guess what you want
“Editorial photo of a woman in her 30s reading in a sun-dappled London park, golden hour, shot on 85mm f/1.4, muted warm tones, --ar 3:4”
Specific, clear, ready to use
Why this works:
The strong prompt specifies the style (editorial photo), subject details (woman in her 30s, reading), setting (sun-dappled London park), lighting (golden hour), camera details (85mm f/1.4), color palette (muted warm tones), and aspect ratio. This removes ambiguity and guides the AI toward a specific aesthetic.
The framework: Step by step
Follow this process to write better image generation prompts every time
Start with the subject: What's the main focus? (e.g., 'product photography of a water bottle')
Add the visual style: Photography, illustration, 3D render, painting? (e.g., 'minimalist product photo')
Describe the lighting: Direction, quality, mood (e.g., 'soft overhead lighting, bright and airy')
Specify camera or medium: Lens, angle, film type (e.g., 'shot on 50mm lens, shallow depth of field')
Set the mood or color palette: Emotional tone, color scheme (e.g., 'warm earthy tones, cozy atmosphere')
End with technical parameters: Aspect ratio, version, quality (e.g., '--ar 4:5 --v 6')