High penetration of distributed photovoltaic causes voltage violation problem in active distribution network. In order to solve this problem, a voltage optimization strategy based on distributed photovoltaic cluster control is proposed for the active distribution network. Firstly, with economic operation as the objective,shiftable load dispatch, the tap positions of on-load tap-changing transformers, and the output of capacitor banks are determined in the day-ahead stage. Subsequently, the approximate voltage sensitivity is calculated based on the day-ahead scheduling results, and clusters are partitioned using the K-means algorithm according to a comprehensive clustering index. Finally, in the intra-day stage, the cluster self-regulation is carried out based on the cluster adjustment characteristics, aiming at minimizing the internal network losses or the node voltage deviations. The inter-cluster coordination optimization is then performed based on the alternating direction multiplier method. Deploy this strategy on the IEEE 33-node system, and conduct a comparative analysis of voltage regulation effectiveness under various weather conditions and different scheduling strategies. The results show that after the day-ahead centralized optimization, the node voltages are within the limit range. After the intra-day rolling optimization, the voltage deviation is further reduced, and the deviation does not exceed 3% on sunny days. The case study results verify that the proposed optimization strategy guarantees voltage quality and meanwhile improves operational efficiency of distribution network.