Craft Midjourney prompts with subject, style, lighting, composition, and parameters.
Midjourney turns text prompts into striking images, but the gap between an average image and a great one is the prompt. Midjourney prompts are built from subject, style, lighting, composition, and parameters - each one a lever you can pull.
The gap between an average Midjourney image and a great one is the prompt - every word either describes the scene or tunes the engine.
The anatomy of a Midjourney prompt: a subject, a style reference, a lighting or mood description, a composition note, and technical parameters separated by double colons. Everything except the subject is optional, but each addition increases control.
Start with the subject, stated clearly: "a red fox in a snowy forest." Then add style: "in the style of a 1980s nature documentary," or "pixel art," or "golden-age oil painting." Style transforms the same subject completely.
Lighting and mood come next: "golden hour, soft rim light," "moody, overcast, cinematic shadows," "neon-soaked, rainy, cyberpunk." These words carry enormous weight in image generation - they set the atmosphere more than any other detail.
Composition words guide the layout: "close-up," "wide shot," "eye-level," "aerial view," "centered subject, negative space." If the default crop is wrong, change the composition before changing the seed.
Parameters tune the technical output. The aspect ratio parameter controls shape, such as a square, a wide cinematic frame, or a vertical poster. Stylization and chaos values control how artistic and how unpredictable the result is.
A complete prompt might read: "a red fox in a snowy forest, golden-hour light, soft rim lighting, close-up, cinematic, in the style of a nature documentary --ar 16:9 --v 6." Every word either describes the scene or tunes the engine.
A complete prompt might read: "a red fox in a snowy forest, golden-hour light, soft rim lighting, close-up, cinematic, in the style of a nature documentary --ar 16:9 --v 6." Every word either describes the scene or tunes the engine.
The classic workflow is generate, then iterate. Generate four options, upscale the strongest, then vary or remix it. Midjourney rewards iteration - the first image is a starting point, not a finished result.
Practice by taking one subject and generating it in three different styles, three lighting conditions, and two aspect ratios. You will quickly learn which words move the image the way you want.
Cramming conflicting styles, ignoring lighting, forgetting aspect ratio, and treating the first generation as final are the usual mistakes.
Generate, upscale, vary, and remix: Midjourney rewards iteration more than any single prompt.