For a long time credit risk assessment has been regarded as one kind of classification issue of pattern i2 dentification , which could not fully meet the requirements of credit risk decision2making. In view of the connotation of credit risk , it is put forward that the probability that credit capital becomes bad debt and the relativity of credit risk should be taken into consideration to assess credit risk. By taking credit risk degree as system output , credit risk assessing and forecasting model based on artificial neural network is established so as to transform assessing mode in effect and provide credit decision2making with more efficient tools and support .