Amid the growing popularity of Computer Vision in Vietnam’s technology landscape, we introduce a jump rope technique recognition application integrated with jump counting using MediaPipe. This application leverages MediaPipe Pose to extract skeletal features from video frames, then applies an LSTM model for real-time jump rope technique recognition. Additionally, the system incorporates a motion tracking algorithm to accurately count the number of jumps. Through experiments on a real-world collected dataset, the application achieves high accuracy in recognizing jump rope techniques and minimal error in jump counting. These results highlight the potential of computer vision in developing intelligent sports training support systems.