Data centers are the backbone of today\'s digital world, powering everything from cloud applications to banking systems and social media. However, running these massive facilities efficiently is a serious challenge they consume enormous amounts of electricity, generate intense heat, and can fail in ways that cost businesses millions. This paper introduces a new system called DT-ML-DCM, which stands for Digital Twin–Machine Learning Data Center Management. Think of it as giving a data center a smart digital brain. A Digital Twin is a live, virtual copy of all the physical machines and systems inside a data center, updated in real time. Machine Learning algorithms then study this virtual model to spot problems before they happen, save energy, and keep services running smoothly. Our framework was tested on a simulated data center with 5,000 servers spread across three cities. The results were striking: energy efficiency improved by 34.7%, faults were detected 41.2% more accurately, unplanned breakdowns fell by 28.3%, and the cost of running the data center dropped by 22.8%. Service agreements with customers were met 99.6% of the time. These findings confirm that combining Digital Twins with Machine Learning is not just a futuristic idea it is a practical, deployable solution that can transform how enterprise data centers are managed today and in the years ahead.