The growing transition of data centers to renewable energy sources necessitates the improvement of energy consumption via predictive analytics. This study outlines a detailed approach to improving energy efficiency in data centers while also promoting the use of renewable energy sources. Employing a comprehensive dataset of 135,873 cases, we undertook a systematic process that included data collection, preprocessing, model training, and evaluation. A range of regression models, including as Cross-Validated Linear Regression, Optimized Ridge Regression, Lasso Regression, Polynomial Regression, Random Forest Regression, and XGBoost Regression, were employed to examine energy consumption patterns in connection to workload measures. A stacking ensemble model was created, utilizing the advantages of top methodologies to enhance prediction accuracy. The stacking ensemble model achieved an impressive R \(^{2}\) value of 0.99997 and an RMSE of 0.00014, significantly enhancing the precision of energy consumption forecasts. This study illustrates that enhancing workload management through sophisticated regression methods results in reduced green energy usage and promotes the sustainable operation of data centers.