Abstract: In order to improve the enthusiasm and flexibility of wind power to participate in the energy/frequency modulation market, a multi-scenario capacity stochastic planning method considering the source-side shared energy storage and the grid connection of electric vehicle combined with wind power clusters is proposed. In accordance with the uncertainty of wind power output, a prediction error persistence probability model is established. Combined the Latin hypercube stratified sampling technology and roulette method, Monte Carlo sampling is controlled to generate a large number of scenes with real wind power characteristics. Then, the operating domain of EV batteries is described, and then the adjustable power and capacity of EV clusters are evaluated. On this basis, an SOC adaptive integration method is proposed to significantly reduce the complexity of scattered control, and a cost-oriented shared energy storage-EV cluster power allocation mechanism based on profit consistency is established. Finally, according to the proposed cooperative operation simulation scenario, the capacity stochastic planning model of shared energy storage and EV cluster based on operation simulation under multiple scenarios is solved. The example results show that the adaptive integration method of EV cluster SOC can be adopted to quickly and effectively integrate a single SOC in the cluster, and can efficiently participate in the source-load interaction. The stochastic programming model considering error persistence can effectively simulate real wind power characteristics and obtain robust capacity allocation, which verifies the effectiveness of the proposed method.