Credit problem is the main bottleneck which hinders the development of e-commerce. This paper analyzes the current representative C2C credit evaluation models of Taobao and Youa, and proposes an improved model. The improved model uses a multi-standard evaluation system and new credit rating rules. And the evaluation algorithm considers the score of multi-standards, category and price of the commodity together, using a weighted system to calculate the credit score then to determine the credit rating. It solves the main problems which the current C2C credit rating algorithms haven't settled. In addition, the model puts forward some identification measures of false trading, the ID verification rules, and a third-party credit certification center, which partly solve the credit fraud problems arising from the credit island and credit speculation. Finally, the paper compares the improved model with current models to show its superiority.