This article aims to assess the existing guarantee fee mechanism of E- Hui by using credit card grading system to charge the various members of the credit rating for different fee ; using three macro factors to forecast the default rate : the amount of credit cards newly issued (lagged 6 months ), annualized 6-month rate of change of composite leading index (lagged 16 months) and the benchmark lending interest rate (lagged 6 months), and the three macro factors are significant. Besides, the bank’s exposure to the amount and the period of actual risk is included in the model. We believe that users do not need to subsidize those users of poor credit and it can introduce more borrowers to improve the imbalance between capital supply and demand in the E- Hui platform through improved guarantee fee system