Virtual Reality (VR) technology has transformed the gaming landscape by offering immersive and interactive experiences. This study delves into the integration of deep learning-based gesture recognition technology into VR magic games, aiming to enrich user engagement and interaction within VR environments by allowing players to manipulate gameplay elements using hand gestures. The research explores a range of techniques and methodologies for implementing gesture recognition in VR games, with a specific focus on leveraging deep learning algorithms for precise and real-time gesture detection. A prototype VR magic game is developed to showcase the viability and efficacy of the proposed approach. Through this game, players can cast spells and execute magical gestures through hand movements, resulting in an engaging and intuitive gaming experience.