Abstract: The increasing integration of renewable energy and the growing impact of climate change have placed higher demands on the accuracy of short-term load forecasting in power systems. Based on historical load data from a provincial power system—including daily 96-point load profiles and meteorological characteristics—this paper proposes an integrated forecasting framework that combines random forest and XGBoost methods with meteorological factors. Experimental results demonstrate that the proposed approach significantly outperforms traditional single-model methods in forecasting accuracy, validating its effectiveness under complex meteorological conditions.