With the continuous and rapid development of our national economy and the improvement of people's income, tourism industry remains rapid and sustained development, which promotes national economy obviously. Since tourism is labor intensive, talents construction has become the core of tourism development. It is a foundational premise for the sustained and healthy development of our national tourism to establish a scientific and applicable forecasting model for tourism talents demand and give the accurate estimate. BP (back propagation neural network) is proved to be an effective technique for classification and forecasting, because of its functions such as nonlinear approach, cosmically parallel disposal etc. In this paper, talents demand model is proposed by analyzing the present situation of Shenzhen tourism industry. After selecting influencing factors, tourism talents quantities forecasting model is established. The model is based on BP neural network theory, which uses time sequence statistical data. At last, through actual data analysis, BP neural network model proves to be a kind of more effectively forecasting model.