Risk perception is an important basis of decision-making for online shopping in consumer to consumer(C2C) E-commerce markets. Firstly, the paper analyzes the influence of seller reputation, product sales and online product reviews on decision-making for online shopping. Then, an evidential network-based dynamic risk assessment model for C2C commodity purchasing is instructed, and the belief rule model is used to reason the network nodes. Finally, the feasibility of the risk assessment model for commodity purchase online is verified by using the multivariate linear regression analysis on the data set from observation information of the E-commerce platform. With the help of this model, consumers can better understand the risk of buying goods in online shopping.