Crises related to water scarcity and air pollution are major problems in the world today. An important issue regarding wasting of the water is surface evaporation from reservoirs. If the surface of the pond is shaded, evaporation is largely prevented. The use of solar panels on surface of water, simultaneously reduces water evaporation and increases the efficiency of electricity generation. In this study, an artificial intelligence method is used to investigate the effect of a floating photovoltaic system on water loss through surface evaporation in a wastewater pond. The nominated studied case is the Yazd wastewater pond near the city of Yazd in central Iran, the water level of which varies for different months. Amounts of evaporation are measured using the Penman-Monteith method. The simulation is performed with all independent variables used as inputs of the neural network, and the pond’s surface area as the dependent variable. After performing sensitivity analysis, the results show that the network with the best structure for predicting water level is the one built with 9 inputs, 35 hidden neurons, and 1 output, which achieve a mean square error of 4.64658E − 20 and correlation coefficient of 0.999. The predictions show that the surface area of the pond will reach its highest level, 322 ha, in November 2025. Between 2021 and 2025, the evaporation reduction due to the PV system will vary from 272.7 ha in January 2021 to 413.9 ha in November 2025. Overall, the floating PV system will reduce evaporation from the pond by up to 70%.