Achieving optimal performance and minimizing injury risk in golf requires precise coordination of body movements. However, self-training in golf often lacks expert feedback, making it challenging for athletes to identify areas needing improvement. This study presents an innovative golf swing evaluation system utilizing computer vision and machine learning to address these challenges. The system focuses on eight key phases of the golf swing: Address, Toe-Up, Mid-Backswing, Top, Mid-Downswing, Impact, Mid-Follow-Through, and Finish. By leveraging the MediaPipe framework, Human Pose Estimation is applied to swing videos, extracting essential body landmarks and joint angles. These features are analyzed using machine learning models-Decision Tree, Random Forest, Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and One-Dimensional Convolutional Neural Networks (1D CNN)-for phase classification. Although the Random Forest model achieved the highest accuracy, the LSTM model was selected for its ability to handle sequential data, ensuring coherent classification of swing phases. The system compares user swings with a professional baseline dataset, generating similarity scores for each phase and joint angle. Detailed feedback is provided to help users improve their swing mechanics. Furthermore, a user-friendly web application supports video uploads, automated analysis, and progress tracking, offering golfers a complementary tool to traditional coaching for self-improvement.