Abstract:Employment confidence index of college student (ECI) can be used as a reference to assist the university administrators to formulate work plans and cope with the current employment situation. So, the most accurate forecasting results are needed, which is directly related to the effect of policy formulation and implementation. In order to solve this problem, the advantages and disadvantages of the model of threshold auto-regression (TAR), back propagation (BP) and gray model were analyzed. According to analysis result, we propose a combined forecast model based on the genetic algorithm (CF-GA), in which GA is used to determinate the weight of each single model. The forecast results show that the performance of CF-GA has a good behavior to other three models, which can be used as an important reference for the university development of relevant policies.