The background of CHB drug expenditure forecast stems from the large population of CHB patients and the high cost of drugs, as well as the complexity of medical resource allocation and medical security policies, making the forecast of CHB drug expenditure an important topic to optimize the allocation of medical resources and reduce the economic burden of patients. In today's society, the error rate of chronic hepatitis B drug expenditure forecast is high, but the accuracy is low, which greatly increases the burden of patients. The SOM algorithm in neural network is an effective drug expenditure forecasting technique. In this paper, the SOM algorithm prediction system is adopted, which greatly improves the accuracy of chronic hepatitis B drug expenditure prediction, and thus the burden of environmental drug expenditure. Finally, the experimental results show that the SOM algorithm is easy to operate and the accuracy rate is 97.89%.