Abstract: Both the traditional time-of-use tariff (TOUT) and the real-time price (RTP) demand response mechanism will generate new load peaks during the low load period. To solve this problem, considering the demand for peak load regulating on the grid side and the different needs and willingness of different users for charging capacity and charging costs, a dynamic optimization method for TOUT was proposed. The proposed method dynamically updateed the peak-to-valley price of each electric vehicle (EV) based on the load information when the EV was connected to the grid, which made up for the shortcomings of the TOUT and RTP charging methods. Based on the proposed dynamic optimization method of TOUT, by establishing a multi-objective function with the most charging capacity and the least charging cost, the particle swarm optimization (PSO) was used to optimize the charging (discharging) behavior of each EV in two stages. And by introducing a virtual state of charge (SOC) to modify the optimized charging (discharging) behavior, each user autonomously responded to realize the coordinated charging (discharging) of the EV. To verify the effectiveness of the proposed method, based on the results of the 2017 National Household Vehicle Survey (NHTS2017), the Monte Carlo (MC) method was used to simulate the charging demand of 1, 000 EVs in a residential area. And the charging demand under different charging strategies, different optimization weights, different participation levels and different V2G (vehicle to grid) responsiveness was simulated and analyzed. The results show that compared with other charging strategies, the proposed optimization strategy can significantly reduce the user's charging cost and the peak-to-valley difference of the load curve.