An optimization algorithm is described for the voltage control of medium voltage distribution networks in presence of distributed generation. Given the current operating point and the forecasted load and generation, the algorithm computes the changes to be requested to the controllable resources in order to ensure fulfillment of technical constraints (voltage at nodes, current in branches) with the lowest overall cost. Distributor’s redispatching costs, modelled by means of binary variables, are minimized while satisfying service security requirements and ensuring service quality, represented by nonlinear constraints. Storage devices are modeled by means of constraints that relate adjacent time periods. The proposed two-step solution procedure is based on decoupling active and reactive variables. In step 1 a MILP model determines the active power production and the use of storage devices that minimize redispatching costs over all periods in the time horizon: in this step a DC network representation is used. In step 2, given the optimal active power productions computed in step 1, reactive variables in each time period are computed by solving an AC optimal power flow model.