Energy storage systems are key components to increase photovoltaic (PV) self-consumption profitability. Indeed, they allow for the intermittency dampening of the PV production so as to adequately cover end-users’ consumption. Given that in most grid-connected PV systems electricity prices are variable, an informed battery scheduling can significantly decrease energy costs. Moreover, energy storage systems can cover consumption peaks to enable contracted power reduction and hence additional savings in electricity bill. This work elaborates on a scalable and flexible optimization system based on production and load forecasting as a Model Predictive Control (MPC) for battery scheduling that aims at minimizing energy costs for consumers. The system provides a 24-hour-ahead battery plan that reduces purchase cost from grid, extends the battery lifetime and guarantees purchases below the maximum contracted power. The formulated problem is solved by means of a MINLP solver and several evolutionary algorithms. Results obtained by these optimization algorithms over real data are promising in terms of cost savings within Spanish electricity market, particularly when compared to the results rendered by other methods from the state of the art. We end by outlying several research directions rooted on the findings reported in this study.