Abstract: The penetration rate of new energy sources such as photovoltaics in microgrids is constantly increasing, and the intermittency and uncertainty bring huge challenges to the distribution network system.Considering the uncertainty of load forecasting in the distribution network, this article utilizes distributed resources within the microgrid based on different time scale control strategies to reduce the impact of uncertainty on the stability of the distribution network.Firstly, aiming at the uncertainty of load and photovoltaic forecasting as well as the different time scale characteristics of demand side loads, a declaration and decision-making plan for the distribution network, microgrid group, and microgrid is established.And a day-ahead and intra-day multi-time scale rolling optimal scheduling strategy is adopted.Then, the day-ahead optimization aims to minimize the operation cost of the distribution network and micro electric network group, and the intra-dayoptimization modifies the day-ahead dispatching volume based on the day-ahead dispatching results to minimize the cost of controllable resource dispatching in the micro grid.Finally, the target cascade method is used to solve the established model.Through numerical examples, it is demonstrated that the proposed method can reduce the peak valley differences and improve the economy of system.