Abstract: The financial institutions concern on the customer credit evaluation. Highly effective and accurate evaluation models can help the financial institutions to reduce the risk and loss. There are several commonly used evaluation models, for example, the logistic regression, decision tree, support vector machine, artificial neural network, etc. We use the same data set to test these models and give the advantages and disadvantages of these models. Among these models, there is not a unified view that which model is the best because eacxh one has their advantages. The logistic regression is the most stable model while the decision tree is the lowest. in stability, MLP-ANN has the better accuracy rate than other models.