Optimal complementarity of source-load resources and coordination of multi-time scales in an Electric Vehicle (EV) charging station can reduce the operational cost and the impact of source-load randomness on system scheduling strategy. A multi-time scale stochastic optimal dispatch model for an EV charging station considering demand response is proposed. In the day-ahead stage, with the minimum daily operating cost as the optimization objective, the conditional value-at-risk (CVaR) is used to measure uncertain risk, and price-based and incentive-based demand response are introduced to optimize the net load. A multi-scenario optimal dispatch model for the charging station considering operational risk constraints is established. On this basis, an intraday rolling optimization and feedback correction control method for EV charging station based on Model Predictive Control (MPC) is proposed. This can reduce the impact of poor net load forecasting accuracy on optimization decision-making. Finally, the feasibility of the proposed model is verified with simulation of an EV charging station, and the impact of demand response and uncertainty risk preference on the operation of the charging station is analyzed. This work is supported by National Key Research and Development Program of China (No. 2018YFB0904500).