Abstract: Wind power penetration in many countries has been increasing rapidly in recent years. Additional energy saving can be obtained by considering wind power with more proper approach in system scheduling and dispatching. Considering wind power output in mid-long term planning coordinated with short-term scheduling can further reduce the fuel consumption by optimizing resource distribution. Wind speed and wind power cannot be predicted more than one week earlier according to the present technology, which means the mid-long term unit commitment considering wind power output much differs from the traditional one and the short-term unit commitment considering wind power. A dynamic clustering idea is presented which determines the commitment plan of some part of conventional units in the mid-long term planning and determines that of the rest units in the short-term planning. It is based on the advantages of the two scheduling periods that mid-long term planning is more conducive to optimize resource distribution and reduce the total start-up costs, whereas short term planning contribute to reduce the load rate and reserve requirement due to the significant reduction in the uncertainty of wind power output. A two-stage monthly unit commitment model considering short term scheduling for systems with wind power integration is developed to minimize the expected monthly total costs. The effective decision variables contain the group of each conventional unit belongs to and the commitment variables of the units belonging to the monthly optimizing group. Simulation results on the revised IEEE-118 bus system verified the proposed method.