Reference evapotranspiration (ETo), a key component of the hydrological cycle, is fundamental for agriculture. Traditionally, ETo is estimated using the Penman-Monteith (PM) method, considered the standard method by the FAO due to its use of multiple climatic variables, providing a solid physical basis. This research aimed to assess machine learning techniques to estimate ETo at the Yauri meteorological station in Peru. Monthly data on air temperature (maximum, average, and minimum), wind speed, relative humidity, and extraterrestrial solar radiation were used. Two machine learning techniques, K-nearest neighbors (KNN) and artificial neural networks (ANN), were trained and tested. To verify their accuracy, scatter plots, box plots, and various performance metrics were employed. These metrics included mean absolute error (MAE), anomaly correlation coefficient (ACC), Nash--Sutcliffe efficiency (NSE), Kling--Gupta efficiency (KGE), and spectral angle (SA). The results indicate that machine learning techniques provide highly accurate estimates and can serve as viable alternatives for estimating ETo, especially in situations with limited meteorological data. The implementation of these methods can significantly improve water resource planning and management. This improvement is particularly valuable in agricultural regions with data scarcity, offering a practical tool for farmers and water managers to make informed decisions and enhance resource efficiency. The integration of machine learning in this context demonstrates its potential to address critical challenges in hydrology and agriculture.