The distributed integration of large-scale and high-density renewables energy complicates the distribution system structure. Besides, the fluctuation and randomness features of renewable energy bring great adverse effect on network power flow and voltage stability. To solve these problems, taking the active power loss and system voltage deviation as the control objectives, the reactive voltage optimization is studied in this paper based on the equipped static var compensator. An improved algorithm combining beetle colony antennae search algorithm with attractive repulsion and bidirectional learning is proposed. On the basis of the above research, together with the enhanced IEEE 33-node distribution system model, both the reactive voltage control effect of the proposed distribution system and the particle swarm optimization algorithm are verified. The results show that the proposed optimization algorithm can optimize the system power loss on the premise of ensuring the system voltage stability. Besides, compared with the conventional particle swarm optimization algorithm, the proposed optimization algorithm has great improvement in calculation time, convergence speed, and optimal search solution.