To effectively address the uncertainty and volatility brought by the large-scale integration of renewable energy into the grid and achieve optimal operation of the distribution network, a dynamic network reconfiguration optimization method considering the time-varying nature of distributed generation (DG) and load demand characteristics is proposed. Firstly, to tackle the non-convexity of the traditional static branch power flow model, slack variables are introduced to transform it into second-order cone constraints, establishing an optimal power flow model based on second-order cone relaxation. Then, considering the constraints of active management elements in the active distribution network environment, a multi-period distribution network reconfiguration model under high penetration of DG is constructed with the optimization objectives of minimizing network losses and electricity purchases while ensuring voltage stability. Finally, the Yalmip toolbox and Gurobi solver are employed for modeling and solving, and the IEEE33-node model is adapted for verification and analysis. The results of the case study demonstrate that the proposed scheme reduces the expected network loss value from 0.176 MW to 0.097 MW, a decrease of 44.89%, and improves the voltage level by 1.40%, thereby validating the effectiveness of the model.