Due to the problem in the roughness of customer segmentation indicator and low accuracy in retail industry, a customer segmentation method in retail industry is propoesed on the basis of clustering analysis of data mining, and a set of RFM based on multiindicator customer segmentation index system is constructed by using entropy value method to give indicator weight and then by using KMeans algorithm to conduct customer segmentation. Empirical research results show that this method is better than the traditional RFM based on segmentation method in the perspective of distinguishing capacity for customer behaviors feature and clustering compactness, and this method, with feasibility and validity, can better solve the problem in customer segmentation in retail industry and improves the customer relation management and marketing decisionmaking quality.