The world’s ever-increasing need for internet access, social networking, and data storage has significantly risen the demand for data center services over the past decades. Data centers as one of the most energy-intensive enterprises require energy-efficient strategies to optimize the IT operations in DC. This thesis presents efficient energy management solutions which primarily focuses on two major energy consuming areas in DC: IT systems and cooling system. The problem of reliability degradation of IT equipment, inappropriate thermal conditions of IT room, inefficient workload placement, and excessive waste of energy has been addressed in this work. Statistical analysis is performed to identify mutually dependent features of different datasets collected from different sources at data center. This study involves application of machine learning models on monitored data for thermal classification of IT room and implementation of deep learning techniques to future forecast resource utilization and energy consumption in DC. A comparative analysis is also conducted with the existing state-of-the-art work to present the novelty and efficiency of the proposed solutions in terms of high prediction accuracy and efficient future forecast analysis. From observations, it is concluded that the solutions proposed in this research provide consistent, effective, and accurate results that can positively contribute to the improvement of energy management in DCs.