In Iran, given the country’s economic structure and due to reasons such as the underdevelopment of capital markets and other non-bank and contractual networks, financing of the real sectors of the economy is mainly undertaken by the banking system. The proper functioning of this system depends on the efficient use of the resources collected. This, in turn, requires accurate assessment of the risks ahead and identification of appropriate ways to deal with these threats. In terms of purpose, the present study is applied research, and in terms of nature and method, it is descriptive-analytical. It aims to model, cluster, and rank corporate customers based on their level of credit risk. In this study, the financial and non-financial indicators affecting customers’ credit risk are first identified through a review of common credit scoring models. After comparing models such as LAPP, P5, and C6, the C6 model is selected as the base model, and the other indicators and models are classified accordingly. To implement the SOM algorithm, map parameters including grid dimensions, number of neurons, initial weight vectors, learning rate, and neighborhood radius are first determined. Then, customer records, each containing quantitative values of financial and non-financial indicators, are presented to the network either simultaneously or sequentially. Accordingly, the final framework of the study uses a combination of entropy-standard deviation weighting, multi-criteria scoring, and unsupervised clustering through the SOM neural network to evaluate, classify, and rank the credit risk of corporate customers. The findings showed that among the initial 29 indicators, 12 financial and non-financial indicators were selected as factors affecting credit risk. Model evaluation identified three clusters as the optimal number. Due to the weak performance of SOM in separating observations, the K-Means algorithm was employed and classified corporate customers into three groups: low-, medium-, and high-risk. Overall, K-Means demonstrated greater efficiency than SOM in ranking customers’ credit risk.