Abstarct: Neuron-muscle coordinated movements are the vital movements of different sport trainings. In these movements a group of muscles is able to perform incredibly delicate and complicated actions under the control and observation of the motor memory and with the aid of other muscles. Through the current sport system, trainings are invariable so that the strong and weak points of individuals are not considered. This causes that the athletes are not able to do the appropriate exercises to improve their weak points. In order to solve this problem, a system is designed and implemented in this thesis which suggests the apt exercise according to each athlete’s abilities. The suggested system consists of a wireless network, including several discrete notes which are controlled by a software baxsed on reinforcement learning algorithm. In this system, the user is asked to react to the activation of each node, such as a node light turning on. The activation time of each node is considered as a reward for a system so that the system has more tendency to reactivate that node in the rest of the training which makes the user’s weak muscles exercise more than other muscles which means it shows more reactions. The result of the process is the improvement of weaker muscle. In this thesis, an athlete is asked to exercise with this device 10 times a day. The results indicate that consuming the manufactured system causes the increase of the reacting rate of an individual’s weak muscle and coordination of his arms and legs.