The volt/var control is an important task undertaken to ensure the economic and safe operation of a distribution network. In order to improve voltage quality, voltage and reactive power control devices are considered, such as on load tap changer, voltage regulators, capacitor banks and distributed generation inverters. Distributed optimization is proposed to solve the ‘meaningless action’ and low efficiency problems in existing models. First, the voltage and reactive power optimization model of the distribution network is established based on the branch power flow equation, which is relaxed to the mixed integer second-order cone programming. Meanwhile, considering the action cost of the devices, a rolling optimization mode based on model predictive control is proposed. The distribution network is divided into groups and distributed optimization is realized based on an alternating direction multiplier method. Finally, a simulation is carried out based on the improved IEEE 33-bus test system. The simulation results show that unnecessary action of the equipment is avoided under the model predictive control strategy. At the same time, the proposed collaborative optimization strategy solves the "dimension disaster" problem and improves the efficiency of voltage and reactive power control in the distribution network. This work is supported by the National Natural Science Foundation of China (No. 51977154).