Irrigation reservoirs are vital for ensuring a stable water supply to nourish crops, but the environmental conditions that influence their geographical location have not been quantitatively determined on a global scale. This study applied a species distribution model (SDM) to predict the locations of irrigation reservoirs based on seven natural and social predictor variables. Under the assumption that the location of an irrigation reservoir reflects the conditions of the downstream beneficiary areas, new social predictor variables were generated to account for the beneficiary grid cells, and the predicted SDM performance was compared to the results of experiments that did not consider beneficiary grid cells. The consideration of beneficiary areas resulted in response curves that were more in accordance with the actual locations of irrigation reservoirs and improved the prediction accuracy of the SDM. The geographical locations of reservoirs were revealed to be most sensitive to social predictors, and the variable importance was improved by integrating information regarding the beneficiary grid cells. These findings highlight the significance of considering the environment surrounding the target grid cell when applying SDMs to water-related infrastructure.