The increasing demand for innovative indoor sports training solutions necessitates the development of effective techniques for action detection in sports. This study presents a novel method for classifying tennis strokes using machine learning algorithms. By leveraging motion sensors such as accelerometers and gyroscopes, the system accurately identifies various tennis strokes, including serves, volleys, backhands, and forehands. Data was collected from multiple participants and annotated for training and testing the models. The Support Vector Machine (SVM) algorithm demonstrated superior performance, achieving an accuracy of 93%, outperforming other classifiers such as Random Forest and k-Nearest Neighbors. This research highlights the potential for integrating advanced sensor technology and machine learning to enhance tennis training, providing real-time feedback and improving player performance. The findings suggest promising applications for indoor sports simulations and coaching tools.