With the rapid development of many Internet e-commerce companies represented by Alibaba, e.g., Jingdong Mall and Pinduoduo in China, the competition, particularly among them for more customers, has intensified. However, the resources of an enterprise are limited, and it is difficult to serve all customers at the same time. Therefore, it is necessary to effectively classify customers and formulate different marketing strategies based on different customer attributes to serve different types of customers in order to maximize the interests of the enterprise. Based on the research results of the customer segmentation and the cluster analysis by domestic and foreign research scholars, this paper selects the consumption data of the company’s customers based on the actual situation of the Chinese e-commerce company, Jingdong Mall, and uses the K-means clustering algorithm, Kohonen neural network clustering algorithm and the two-step clustering algorithm, as well as the improved K-means clustering algorithm, and other four clustering methods respectively conducted cluster analysis on the company’s customers. Combined with the actual situation of the company, the graphical method and various test methods were used to test the clustering effect. The results show that as compared to the other three algorithms, the two-step clustering algorithm exhibits a better performance in the actual situation and the theoretical test. Then, the customers are divided into five categories, each category is separately analyzed, the characteristics and differences of each category of customers are determined, and different marketing strategies are formulated for the customer groups with different characteristics to provide suggestions for enterprises to differentiate marketing and management in order to achieve the maximum benefits with cost savings.