AbstractThis paper aims to help financial institutions identify credit risk and reduce default losses more effectively by improving the accuracy and efficiency of the credit evaluation process using two methods. The paper first mines indicators that comprehensively describe the supply chain aspects of credit risk to obtain a precise credit risk portrait of the enterprise. The study then innovatively designs a random forest‐weighted naive Bayes (RF‐WNB) model that offers good interpretability and generalization qualities. The model forms an effective two‐stage process of enterprise credit risk evaluation that selects key features, quantifies the importance of each, and then evaluates credit risk. By applying empirical analysis to 1363 listed enterprises and performing one typical case analysis, we verified the effectiveness of the indicators and the two‐stage RF‐WNB model in evaluating enterprise credit risk.