It is important to consider the operational characteristics of wind power in the generator maintenance scheduling for optimizing the use of wind power and ensuring the safety of system operation. A comprehensive model of generator maintenance scheduling， considering the operational characteristics of wind power， the fluctuation characteristics of load and the regulating and maintenance characteristics of units， is established. The model optimization goal is the combination of system maintenance cost and revenue loss of wind energy， which is mainly constrained by the power generation reliability and peak-load regulating adequacy. Based on Benders decomposition， the model is decomposed into two parts: a master problem（deterministic multi-objective integer programming） and three sub-problems（revenue loss of wind energy calculation， power generation reliability and peak-load regulating adequacy） for decentralized coordination， and this method is applied to make maintenance scheduling for all the generators in IEEE-RTS test system with large-scale wind power integration. The results prove that the proposed method is feasible and effective.