Effective customer classification is a key issue for logistics enterprises to carry out customer relationship management, which has been concerned and studied by relevant institutions. Based on BP and AdaBoost algorithm, a new customer classification method for logistics enterprises is proposed. First of all, the logistics customer value evaluation index system is designed, and the different customer characteristics are analysed. Secondly, in order to solve the problem of low accuracy of the traditional BP network, the AdaBoost algorithm is used to linearly combine the traditional BP network to produce a strong classifier. The simulation results show that the application of the new customer value classification method based on BP-AdaBoost in customer classification improves the classification accuracy.