This study presents a new way to improve technical training in para-athletic disciplines (for athletes with musculoskeletal disorders). The experimental group included 5 men and 1 woman. 5 athletes were preparing to perform in shot put, 2 athletes were preparing to perform in discus throw and 5 athletes were preparing to perform in javelin throw. An algorithm for technical training was developed, comprising three stages. The first stage includes high-speed video recording of the performed exercises in the front plane (from the front, from the back), sagittal plane (from the right, from the left), and transverse plane using a video camera attached to the quadcopter. The second stage includes synchronization of video recordings, recognition of markers, and obtaining 3D biomechanical model of the exercise. The third stage includes the analysis of kinematic characteristics of equipment of competitive exercises with the subsequent interpretation and comparison with the model characteristics revealed earlier by identification of parameters of available motor actions. The recommendations for updating the individual locomotion in the activity and guidance on the selection of major exercises were elaborated. According to the developed algorithm, six measurements were performed in each competitive discipline during the preparatory period of 2 months. After identifying the reserves of equipment, the technique was corrected in each competitive exercise, and the individual recommendations for optimizing the technique were given. At the end of the experiment, the average improvements of athletes were 0.85±0.13 m in discus throw (n=5, P<0.01), 3.15±0.69 m in shot put (n=2, P<0.05); 2.53±0.41 m in javelin throw (n=5, P<0.001). At important competitions at the end of the experiment, the athletes set 10 personal records.