Rehabilitation for upper arm and wrist problems is very important for individuals who rely on their arms for daily activities and work. Traditional methods are often monotonous, resulting in decreased engagement. Our study introduces an immersive and interactive exergame-based training environment designed to improve wrist mobility with real-time progress tracking. It involves an interface between the six-axis IMU sensor and Arduino board to capture and quantify wrist movement data, which is integrated into unity-based games for training. A wristband controller is custom-designed for this purpose. Two training environments—underwater flappy bird and endless forest runner—were developed with varied levels to facilitate wrist flexion/extension and radial/ulnar deviation, respectively. The user was thus able to navigate through both the training environments by moving their wrist in the appropriate directions. Further, a gesture recognition model was developed using computer vision techniques to validate the project by comparing the model’s output with real-time sensor data. The trial was conducted on 10 subjects who were found to have spent over 11 h per day using computers. Statistical analysis was conducted to assess the model’s accuracy and effectiveness. The system’s usability was also evaluated by user testing and feedback collection. Movement data was compared with the gesture data obtained using computer vision techniques. The results were analysed using statistical parameters such as accuracy and precision. An aggregated accuracy of over 81% was obtained. Through the fusion of gaming elements with therapeutic exercises, this project offered a novel immersive method of rehabilitation.