Abstract: Electric vehicle（EV） charging demand estimation is an important precondition for studying the vehicle-to-grid（V2 G）interaction. Therefore, this paper proposes a charging demand prediction model of EVs driven by driving trajectory data,constructs a decision-making model of users to choose to participate in V2 G response by further considering the multi-dimensional benefits of users, and analyzes the regulation potential of regional V2 G response capabilities. Firstly, the big data set of driving trajectory is cleaned and mined, and a prediction model for the spatio-temporal distribution of EV charging demand is constructed based on the dynamic energy consumption theory. Secondly, based on the social behavior theory and considering the electricity demand utility, economic utility, environmental protection utility and social utility, the probabilistic selection model of EV users participating in V2 G response is constructed. The model not only considers the heterogeneity of EV users, but also reflects the interactive influence of user decisions. Finally, a V2 G responsive capacity regulation model is established to analyze the adjustment effect of V2 G responsive resources on the regional load. The results show that the proposed model can not only effectively estimate the spatio-temporal distribution characteristics of EV charging demand in a certain urban area, but also obtain the number of potential EV users who choose to participate in V2 G response in this area, which provides support for studying the regulation potential of V2 G responsive resources on the regional load.