Groundwater salinization is a growing threat to water quality, especially in water-stressed regions where natural salinity sources, climate variability, and intensive land use interact. Here, we examine salinization mechanisms across Israel’s seven hydrological basins, focusing on four representative basins that span the main climatic, hydrogeological, and land-use gradients. Long-term chloride records (1998–2022) were integrated with spatial datasets representing physical, climatic, and anthropogenic controls. Six machine learning models were evaluated, with Extreme Gradient Boosting and a temporal Convolutional Neural Network providing the strongest predictive performance. SHapley Additive exPlanations were used to interpret basin-specific controls.,Results show that dominant controls vary among basins. In the Coastal Aquifer, inland saline groundwater bodies were the primary control, while river corridors acted as secondary contaminant and solute-transport pathways, indicating structurally driven salinization amplified by anthropogenic pressures. In the Yarkon–Taninim karstic carbonate aquifer, reduced rainfall was associated with pronounced salinity increases, reflecting the rapid hydraulic response of the karstic system, where reduced recharge lowers hydraulic heads and promotes the upward migration and mixing of deeper saline groundwater with overlying freshwater. In the Sea of Galilee Basin, salinity patterns were associated with hydroclimatic and landscape controls, including temperature, topographic wetness, recharge-related variables, and localized surface–groundwater connectivity. In the hyper-arid Negev–Arava basin, salinity was governed primarily by regional saline groundwater bodies and the aridity gradient, while agricultural effects were secondary and spatially dependent.,These findings demonstrate that similar predictors can represent different salinization mechanisms across contrasting groundwater systems, highlighting the need for basin-specific interpretation in management and showing how explainable machine learning can support process-based understanding of salinization.