According to the problem of image similarity comparison, an image similarity comparison method based on data field clustering is proposed. In the method, at first, image information is converted from RGB color model to HSV color model. And then, each pixel point is projected into the HSV 3D space, the main color is extracted through data field clustering algorithm. At last, image similarity judgment is performed according to the main colors of the two pictures. Experimental results show that the algorithm has better similarity judgment effect.