Abstract: With the integration of large-scale distributed generation, the active distribution network encounters great challenges in dealing with randomness of the distributed generation output and coordinating various reactive power resources in the network. In this paper we propose a multi-time scale reactive power/voltage optimization control method for active distribution networks based on the combination of day-ahead improved multi-objective grey wolf algorithm and intra-day second-order cone programming method. Firstly, considering the regulation characteristics of various adjustable reactive power resources in the network, a day-ahead reactive power/voltage optimization control model is established with the goal of minimizing network loss, voltage deviation and operation cost of the discrete voltage regulating equipment. An improved multi-objective grey wolf algorithm is presented to address the given model. Secondly, the dispatching requirements of the active distribution network in the intra-day stage are taken into account. Combined with the rapid reactive power regulation capability of the distributed generation, a second-order cone programming model is established aiming at minimizing both network loss and voltage deviation. Finally, the proposed method is verified on the IEEE 33-bus example, demonstrating its capability to accommodate different optimization objectives while exhibiting excellent convergence and real-time performance.