A customer segmentation method based on dynamic SOM and RFM indicators is presented. Firstly, customer clusters is generated by DSOM clustering analysis. Secondly, customer loyalty, customer scale and customer credit are simulated by RFM behaviors of customer, and then customer clusters are identified. Finally, the value of customers can be decided according to the customer segmentation results, and enterprises are able to provide better different service levels to different customers. It provides customer strategies of enterprises with an effective support and decision-making.