Abstract: Aiming at the coordination problem between system optimization scheduling and local distributed control in AC/DC hybrid distribution network, a centralized-distributed control architecture was proposed. At the centralized control layer, the minimum network loss and the maximum level of new energy consumption were optimized. The optimal power reference value of each controllable unit in the system was calculated by optimizing the scheduling algorithm to provide an initial operation instruction for the local control layer; in the distributed control layer, the DC bus voltage was divided into normal state, risk state, and limit state. In the normal state and risk state, each slave converter station realized the smooth operation of the system through the droop control based on the optimized dispatching command. When the DC voltage exceeded the limit, the battery acted as a backup unit for charging and discharging control to realize secondary recovery of the DC voltage. The simulation results based on Matlab/Simulink show that the control strategy can flexibly and efficiently adjust the system power under normal operating conditions. When the main converter station exits, the system is realized by charging and discharging of the battery unit which can improve the reliability of the distribution network.