A double time scale scheduling strategy based on multi-agent technology is proposed for microgrid community system with renewable energy uncertainties, which includes day-ahead stage and real-time scheduling. In each particular scene of day-ahead stage, DispatchAgent uses the Monte Carlo simulation to generate scenario. By introducing the mutual energy transfer model between microgrids, the corresponding energy scheduling matrix is calculated. According to the economic operation of microgrid community, PSOAgent uses particle swarm algorithm to get the day-ahead scheduling under the scenario. In real-time scheduling, the output data of t moment of new energy are received by MGiAgent and compared with forecasted expected value, and it works out real time scheduling scheme of t moment quickly to meet the demand and supply balance of t moment of microgrid i and achieve cost optimum based on the expected day-ahead scheduling scheme, thereby realizing the economic operation of microgrid community. Compared with traditional time scale method, the proposed method has smaller scheduling bias than the traditional method in real-time scheduling, and can effectively reduce the abandon and exhaust phenomena in microgrid communities. This work is supported by National Natural Science Foundation of China (No.61673174) and the “111” Project (No. B17017).