This thesis aimed at the science and technology?SME, firstly established a multi-level evaluation index system of credit risk , and use the t test and factor analysis?to simplify the index system , then on the basis of the index system establishes a credit risk evaluation model of the science and technology?SME based on the MLP neural network. Finally, use 66 high-tech listed companies as the research sample for empirical analysis , the test results show that the model established in this paper for small and mid-sized enterprise credit risk evaluation model has a high accuracy.