The swift expansion of solar photovoltaic (PV) technology has introduced significant challenges for those overseeing electricity distributions due to its reliance on weather conditions. In this study, predictive modeling techniques including Multilayer Perceptron (MLP) Regression, Gradient Boosting Regression, and Extreme Gradient Boosting (XGB) Regression were employed to enhance solar PV output forecasting. Comparative analysis utilizing performance metrics such as Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and R-square revealed that the XGB model outperforms others. Specifically, the XGB model demonstrated an MAE of 0.0653, RMSE of 0.10461, and R-squared of 88.66%, showcasing its superior accuracy and efficiency in solar power management. This synergy between machine learning and energy efficiency heralds a promising direction towards achieving a more sustainable, decarbonized, and digitalized energy sector.