Exercise using a punching bag is very useful for the body, as long as it is done according to the individual’s abilities and not excessively. So far, punching bags have not been equipped with a level identifier of the force of the punches performed. The purpose of this study is to propose a smart punching bag to identify the force level of the punches that have been performed using fuzzy. The method applied in this study is fuzzy mamdani, with triangle function membership, and defuzzification using centroid and bisector. The input of the system is the result of the FSR 402 sensor and accelerometer, while the output of the system is the level of the punches that have been performed. Based on the experiments that have been carried out, the proposed approach punching bag for punch force level identification in self-defense training using fuzzy logic can be implemented, although using different defuzzification parameters produces different output results. However, this difference is not too large, namely the average value of the delta difference between the two experiments is −0.01175, the minimum value is −0.388, and the maximum value is 0.3645. This research implies that a monitoring system approach is obtained from punches hitting the punching bag. The contribution of this research is expected to help the training process when doing punching bag hitting exercises. When doing a punch, later it can be known how much force hits the punching bag.