An image encoding/decoding method and apparatus, in which at least one of a plurality of color component images except a first color component image is predicted from motion estimation information or a reconstructed color component image of the first color component images using correlation among the plurality of color component images. The image encoding method includes performing predictive-encoding on the first color component image among the plurality of color component images forming a single image using a general encoding method, and performing motion estimation and compensation on the at least one of the other color component images or predicting the at least one of the other color component images using a reconstructed first color component image.