Under the limited financial market scale, the loans and credit business of banks is quite competitive in Taiwan. Therefore, the quality of client credit is a critical issue that banks pay considerable attention to. This research employs logistic regression model to explore the main default risk variables that need to be considered when banks verify consumer loans, and then builds a credit scoring model as a standard that can be referenced. The empirical results are as follows. Firstly, both the appliers with compensatory behavior and have been queried by banks on JCIC have the highest default risk. Secondly, default risk is lower if appliers with higher total unsecured debts, or their education level is between senior high school and master degree, and so do the married appliers. This study can be a reference for banks while verify small amount credit loans.