The invention provides an image retrieval system and method based on color feature extraction. The image retrieval method retains the advantage of invariance during rotation and zooming in an existing retrieval method, partial color features for reflecting space distribution are added, and the capacity of reflecting color space distribution information is achieved. The image retrieval method includes the steps of firstly, extracting color edges of images to obtain an image edge matrix, and carrying out retrieval on largest color blocks and color blocks adjacent to the largest color blocks in the images based on the image edge matrix; secondly, extracting color feature pairs of the color blocks through color space changes, and measuring the similarity between the images with the Euclidean distance. In addition, rapid retrieval of the images can be achieved through the image retrieval system. According to the image retrieval system and method, the defect that in the prior art, space information is lost due to the fact that the numbers of same gray values of histograms based on colors are compared without reflecting color space distribution is overcome, the image retrieval precision ratio and the image retrieval recall ratio are accordingly and effectively increased, and the image retrieval flexibility is improved.