The fluctuation is an important factor that limits large-scale integration of wind power into power grid. In order to improve penetration level of wind power, the EVs based on V2G are considered to participate in regulating wind power while considering charge-discharge characteristics of EV battery. Thus, in this paper, an optimized EV charge-discharge control model is proposed to reduce output fluctuation of wind power. The Monte-Carlo method is used to simulate the stochastic wind speed based on Weibull probability density function. Finally, Genetic Algorithm (GA) is adopted to solve the problem. Results indicate that the EVs based on V2G can reduce the wind power fluctuation level to some extent, absorbing the wind power surplus and compensating the of wind power shortage.