The existing human movement intensity monitoring only rely on a single physiological data changes, with general standards to determine whether the intensity is too large, not only the lack of flexibility in monitoring, and monitoring accuracy and real-time poor. In order to monitor exercise intensity more conveniently and accurately and improve the curve existing in the above method, a human exercise intensity monitoring method based on smart wristband was studied. For PPG signal collected by smart bracelet, the artifact removal process is carried out by GAN After extracting PPG and ECG signal characteristics, human body motion posture was identified by combining the motion posture change data collected by the bracelet. According to the heart rate and load intensity range of standard setting is roughly, based on the SVM model to estimate abnormal physiological data of firefly algorithm, realization of human movement monitoring intensity. The experiment was carried out among 100 monitoring objects, and the accuracy rate of monitoring exercise intensity by using smart bracelet was higher than 97%, and the monitoring effect was better by giving feedback to the change of intensity of the monitoring object in real time.