Abstract Energy from the sun is weather-dependent. In modern electric grids that is a shortcoming; generation (and load) has to be regulated accordingly. This issue is a cornerstone for an effective transition to a renewable-based energy system. Weather forecast algorithms can predict photovoltaic production but, in real life conditions, their reliability is only partially effective with respect to the actual grid operation requirements. In the paper, Energy Storage Systems are adopted to compensate the mismatch between the injections of a photovoltaic power plant and the day-ahead market power schedule: the final goal is to achieve the full programmability of the photovoltaic resource by minimizing energy imbalances, as defined in the Italian regulatory framework, on an hourly basis. In particular, the optimal design of the storage apparatus (nominal power and capacity) is defined according to the regulating performances required. Moreover, three forecast models are tested to evaluate the impact of weather prediction accuracy on the ESS design. Finally, the benefit/cost ratio of the ESS application is assessed according to the main economic and technical parameters (ESS cost, round trip efficiency, lifespan). The analyses are performed on data measured on a real power plant, with hypotheses consistent with the actual Italian scenario.