Monitoring how much water trees need is key to managing irrigation and improving drought resilience in orchards. Stem water potential (SWP) is a useful indicator of tree water stress, but measuring it everywhere in an orchard is time-consuming and expensive. In this study, we used machine learning to predict SWP across a walnut orchard by combining field measurements with weather data, soil and plant information, and satellite imagery. We tested four models—Random Forest, Support Vector Machine, Gaussian Process Regression, and XGBoost—along with an ensemble model that combines all of them.