To scientifically assess water resources value in the Yellow River Basin and support refined management and high-quality development, from a dynamic perspective, this study used a value index to represent water resources value and analyzed its spatiotemporal evolution trend. By applying the Geodetector (GD) to identify driving factors and using genetic algorithm (GA) to optimize key parameter of the XGBoost model, a GD GA XGBoost model was constructed and used to predict water resources value in the Yellow River Basin from 2022 to 2026. The results show that water resources value exhibits significant spatiotemporal differentiation. Temporally, it follows an inverted U shaped trend characterized by growth, stabilization, and then declining. Spatially, upstream areas maintains consistently high water resources value, dominated by high high clustering; midstream areas first rise and then decline, with diverse clustering patterns; downstream areas generally have relatively low water resources value, primarily characterized by low low clustering. Water resources supply factors, such as annual precipitation, mean elevation, and mean slope, consistently play dominant roles. While demand side factors initially show limited explanatory power but gradually increase in influence, with multi factor interactions undergoing dynamical changes. Predictions indicate an overall downward trend in water resources value from 2022 to 2026, suggesting potential alleviation of future water resources supply demand conflicts.