Data centers are major contributors to global energy consumption, with significant implications for operational costs and environmental sustainability. As energy demand increases, optimizing energy usage within these facilities has become essential. This study investigates the application of machine learning-based predictive analytics to enhance energy efficiency in data centers. By leveraging models such as Random Forest, Neural Networks, and Deep Learning, predictive analytics forecasts energy demands based on variables like temperature, workload, and time of day. Results from multiple case studies reveal that machine learning models can reduce energy consumption by up to 20%, offering a sustainable solution without compromising data center performance.