How to Write Better Midjourney Prompts

Effective Midjourney prompts combine specific descriptive language with strategic parameter usage to generate precise AI images. Better prompts reduce iterations and produce images that match your creative vision.

1. Start with a clear subject description. Begin your prompt with the main subject using specific nouns rather than generic terms. Write 'golden retriever puppy' instead of 'cute dog' or 'Victorian mansion' instead of 'old house'. Place the most important elements first since Midjourney weights earlier words more heavily.
2. Add descriptive details and style modifiers. Include lighting conditions, materials, colors, and artistic styles after your main subject. Specify elements like 'dramatic lighting,' 'oil painting style,' or 'shot on 35mm film.' Layer these modifiers to build complexity without creating conflicting instructions.
3. Set aspect ratio with --ar parameter. Add --ar followed by your desired ratio at the end of your prompt. Use --ar 16:9 for landscape, --ar 9:16 for portrait, or --ar 1:1 for square images. This parameter controls the canvas dimensions before generation begins.
4. Control image quality and detail. Use --q 1 for faster, standard quality or --q 2 for higher detail and longer processing. Add --stylize followed by a number from 0-1000 to control how much Midjourney applies its artistic interpretation. Lower values stay closer to your literal description.
5. Use negative prompts to exclude unwanted elements. Add --no followed by elements you want to avoid, such as --no text, hands, or watermark. This parameter helps eliminate common AI artifacts or unwanted objects that frequently appear in generated images.
6. Structure complex prompts with weight brackets. Use double colons :: to separate distinct concepts and prevent them from blending together. Apply weight modifiers with numbers after double colons, like 'red car::2 blue sky::1' to emphasize the car over the sky. Weights above 1 increase importance.
7. Iterate and refine based on results. Generate your initial image, then use the V buttons to create variations or U buttons to upscale promising results. Modify your prompt based on what worked and what didn't, adjusting descriptive language and parameters systematically.