Abstract: A novel image retrieval method based on color and texture feature is proposed.Based on the quantization of the color in HSV model，color histogram is introduced as the color feature.Then，the quantization results are classified into different status and the transition probability matrix is presented to describe the texture feature.After that，color relative entropy is given as a texture descriptor based on entropy theory and the state relative histogram.Finally the color histogram and color relative entropy are combined together as an index for image retrieval.Experimental results show that the new method is efficient and it provides noticeable improvement to the performance of image retrieval.