The demand for physical fitness and mental health has led to the substantial popularity of martial arts and combat sports. Research on training tracking devices has also been increasingly interesting due to the challenges of boxing, such as the risks of injuries and the need for correct training techniques. The aim of this research is to design a device to monitor boxing training to support athletes and coaches. The proposed device is integrated with an accelerometer and a gyroscope and is attached to the punching bag. The linear and angular accelerations due to the punches are measured by the device and are sent to a computer for further analysis. Another circuitry is designed with a load cell to measure the true punching force. An artificial neural network (ANN) model is developed to learn the true punching force from the punch-induced bag accelerations. As a result, a high correlation of 0.99 of the punching force between the true and the predicted values indicates the feasibility of the proposed device for estimating the punching force based on the punching bag accelerations. Furthermore, a multi-function user interface is designed to allow athletes and coaches to monitor and assess their training activities both in real time and in history.