Risk assessment is an important technology to ensure the security of information system. Information system risk assessment is subject to a number of effects and has dynamic and changeable, in order to improve the accuracy of information system risk assessment, this paper puts forward a information system risk assessment model based on hierarchical analysis method and neural network. At first, the current situation of the information system risk are analyzed, and this paper builds the evaluation index system according to the experts, and then the analytic hierarchy process is used to estimate the weights of evaluation indexes, and SVM neural network is used to establish the information system risk assessment model, at last, the reliability of the evaluation results is tested by using the risk data of information system. The results show that the proposed model can reasonably determine the weight of evaluation index, it can describe relationship between risk factors and expected value of information system, obtains ideal assessment results, and so it has certain practical value.