Public hospitals are a fundamental service industry in China, and the safe and sustainable operation is crucial for the stability of the nation and society. The production and operation activities of public hospitals havee many uncertainties, requiring effective prevention, control, and early warning of financial risks. To this end, research is conducted on the construction method of a financial risk warning model for public hospitals based on cluster analysis. Starting from the four perspectives of growth, it selects early warning indicators of financial risk of public hospitals from financing, investment, operation. According to the linear relationship between each index, the entropy weight method is used to determine the financial risk index weight. The index confidence was obtained based on cluster analysis, and the standard value of financial risk warning was calculated. The neural network algorithm is used to build a risk early warning model, score the financial data in public hospitals, and realize the financial risk division of the data. The experimental results show that the model can be used to classify various financial risks of public hospitals, the prediction accuracy is high, and it has the pre-withholding warning function.