In the field of sports analytics and training, exploiting advanced technologies such as Computer Vision and Machine Learning has become increasingly prevalent. This paper presents a novel approach to player analysis and sports training, focusing particularly on racket sports such as badminton and tennis. Using tools such as OpenCV, Google’s Mediapipe for the extraction of body coordinates and algorithms like You Only Look Once (YOLO) for object detection, MIL tracker and bounding boxes for the selected player movement tracking, the proposed model aims to provide comprehensive insights into player performance and technique. Sports training is provided with the help of Sequential Deep Learning Model (FeedForward Neural Network). This model is trained on body coordinate data extracted from videos of coaches demonstrating specific shots. By analyzing these coordinates and providing feedback on player posture and technique, the model facilitates individualized coaching experiences without the constant need for a physical coach. Through techniques such as individual player time detection using YOLO and DeepSORT, the model provides insights into player activity during the game, assisting in performance optimization and strategic decision-making.