Abstract This paper proposes a distributed voltage control (DVC) scheme for smart distribution networks with high penetration of inverter-based distributed generators (DGs), aiming to optimally coordinate DG units and on load tap changer (OLTC) transformer to regulate the voltages within the feasible range. The proposed scheme consists of two important parts: (1) distributed information synchronization (DIS) framework and (2) distributed model predictive control (DMPC)-based voltage control scheme. The DIS framework is established based on the consensus protocols to synchronize the specific information about the critical bus voltages and potential OLTC actions. The DMPC-based voltage control scheme is presented, in which each DG unit only exchanges information with its immediate neighbors and solves the local optimal control problem. Two control modes are designed to better deal with different operating conditions. In the normal mode, only the reactive power outputs of DG units are optimized to mitigate the voltage deviations. In the corrective mode, both of the active and reactive power outputs of DG units are optimally controlled to correct the severe voltage deviations. To mitigate the mutual interaction between the DGs and OLTC, the potential actions of OLTC are predicted and considered in the optimization problem of each units. The control performance of the proposed scheme was demonstrated using a real medium-voltage (MV) distribution network with two feeders under both normal and large-disturbance conditions.