To improve the local accommodation of distributed renewable energy and realize hierarchical optimal scheduling model for a distribution network, this paper proposes a multi-objective ant colony dynamic partitioning algorithm based on MOEA/D and a day-ahead optimal dispatching model based on dynamic partitioning. Using a power flow tracing algorithm and bipartite modularity in complex network theory, an energy bipartite modularity index that quantifies the degree of energy coupling between partitions is proposed. Based on a Jacobian matrix of power flow calculation, heuristic information in the ant colony algorithm is derived. Combined with the prediction scenarios, the energy bipartite modularity and power reserve of the partitions are used as the objective function, and the multi-objective ant colony algorithm is used to generate dynamic partitions. A day-ahead optimal scheduling model based on dynamic partitions is established with the objectives of the partitions' communication line power, insufficient flexibility rate and lowest cost. The Pareto optimum is determined based on the NSGA-II algorithm. Finally, based on the IEEE33 bus distribution network, the proposed model and method are verified. The results show that the dynamic partitioning and day-ahead scheduling using this method can effectively improve the system's ability to deal with the uncertainty of renewable energy, and lay the foundation for suppressing the fluctuation of renewable energy locally. This work is supported by the Natural Science Foundation of Fujian Province (No. 2017J01480).