This chapter presents an energy management method designed to optimize electricity consumption by mitigating peak demand through the strategic scheduling of an irrigation pumping system’s operating hours on an isolated island. Initially, Machine Learning (ML) models are employed to accurately predict island-wide electricity consumption and hourly energy production from Renewable Energy Sources (RES). Subsequently, these predictions are integrated into a heuristic optimization algorithm that establishes the optimal pump operating schedule, thereby minimizing consumption peaks while adhering to operational constraints. The efficacy of the proposed method is evaluated using the case study of Tilos, a remote Greek island equipped with an energy management system that oversees and regulates local water pumping stations for household water supply and irrigation. The findings suggest that intelligent pump scheduling in a small-scale island context can decrease daily and weekly electricity consumption variability by over 15% without incurring any associated monetary costs. Furthermore, this chapter concludes that the potential advantages of this approach are contingent upon three critical factors: the volume of load that can be shifted daily, the accuracy of the forecasts, and the extent of electricity generated by photovoltaic (PV) systems.