Micro-hydropower has been highlighted as a potential technology suitable for installation in irrigation networks to reduce system overpressures and to reduce the net energy consumption of the irrigation process. However, the full impact of this technology on a large regional scale is unknown. Artificial Neural Networks and regression models were used in this research to predict the energy recovery potential for micro-hydropower in on-demand pressurised irrigation networks across a large spatial scale. Predictors of energy recovery potential across spatial unit areas included: Irrigated land surface area, irrigation crop water requirements, rainfall, evapotranspiration, and mean topographical slope. The model was used to predict the energy recovery potential across the 164,000 ha of the Spanish provinces of Seville and Cordoba in the absence of hydraulic models. A total of 21.05 GWh was identified as the energy potential which could have been recovered using micro-hydropower during the 2018 irrigation season. This amount of energy would have potentially reduced the energy consumption of the irrigation process in this region by approximately 12.8%. A reduction in energy consumption in the agriculture sector of this magnitude could have significant impacts on food production and climate change. The main novelty of this paper lies in the assessment of micro hydropower resources in operating irrigation networks on a large geographical scale, in areas where no information is available. It provides an approximation of the existing potential using computational methods. (C) 2020 Elsevier Ltd. All rights reserved.