In today’s diversified development model, the combination of modeling and business decision development is particularly important. The advanced theoretical business model established by modeling enables more efficient and accurate financial analysis. In the original enterprise profit evaluation model, the DuPont analysis method cannot take into account the development capability of the enterprise very well. This article takes Tencent Music as an example, and improves it on the basis of DuPont analysis. The Enterprise Capital Profit Model was proposed. At the same time, the LASSO regression based on cluster analysis is used to screen, analyze, and diagnose the financial data of Tencent Music in recent years, which verifies the validity and feasibility of the model. This paper uses the report data combined with statistical modeling to optimize the traditional financial evaluation method of enterprises, better find problems, and provide strategies for the further development of enterprises. Likewise, the method can be extended to other businesses to help them analyze their financial situation and provide a reference for future development.