This paper utilizes the SOA framework to realize the system structure design, combined with the total flow of the claim processing, the design of the acceptance and audit processing, responsibility determination and reminder management of 4 modules, constituting the customer claim management system as a whole. The improved LEACH algorithm is utilized to control the network topology, and the cluster tree combined with data fusion is used to reduce the energy loss of the nodes in the network. Based on the weighted queuing scheduling algorithm, the feedback processing module is designed to determine the grouping rules and calculation methods, and the decision tree algorithm is used to build the demand management prediction model. The accuracy of the prediction of the decision tree algorithm is verified to realize the case information management statistics. Analyze the utility of the improved LEACH algorithm and test the system’s performance. The experimental results show that the improved LEACH algorithm is more balanced in the range of [0,100] dead nodes, and the energy consumption is well balanced. The total delay is 16.13-time slots, which is the lowest delay among the four algorithms. The node death time is 1045 rounds, having the longest network life cycle. The system performance is tested using IO and memory-intensive loads. Before writing a file of 16GB, the system time responsiveness stays around 80%, and the system CPU utilization and disk IO utilization increase as the file is written. The system performance is good.