Abstract This paper proposes a real-time energy management strategy for pumped hydro storage systems in farmhouses to manage surplus renewable energy. The proposed system meets both electricity and water demand in a farm. The novelty of this paper is its combination of a scheduling method and a real-time controller to take into account both present and future conditions of the microgrid. The scheduling part determines irrigation times, required stored water, and pumped hydro storage schedule. The real-time controller receives the schedule and current condition of the microgrid in order to adjust the pump power and turbine flow rate efficiently. Two methods of fuzzy logic and artificial neural network are tested to investigate which can address the forecast error problem more economically. An innovative approach is presented to produce target data for artificial neural network training. The designed system is simulated for 365 days to investigate the effect of real-time management on the performance of the microgrid on both sunny and cloudy days. The proposed energy management system is applied in an experimental setup, tested with a real pump and turbine. Results show that a real-time management system could keep the stored water level the same as the scheduling method; however, the pump and turbine can be controlled more cost-effectively. Finally, an economic study is conducted to determine the payback period of the system.