Optimal management of micro-grids requires anticipating the supply-demand unbalance. This work aims at developing a method to integrate real day-ahead deterministic forecasts of photovoltaic (PV) production and of system loads in the management of an ESS integrated inside a micro-grid. Dynamic Programming (DP) has been chosen to optimize the cost of the micro-grid operation. To test the developed method, a real educational Net Zero Energy Building equipped with a PV roof is considered. Compared to a management that does not take advantage of forecasts, the developed method allows decreasing the operating cost of the system.