Microsurgery enables precise manipulation of small and complex anatomical structures. However, it is a physically and mentally demanding surgical technique. Advances in wearable technology and artificial intelligence (AI) have improved the development of innovative solutions in this field. The objective of this study is to evaluate the surgeon’s physiological and ergonomic parameters by using EMG wearable technology during performance in conventional and robot-assisted microsurgery (RAM). In addition, this study seeks to develop predictive models for musculoskeletal risks, such as localized muscle fatigue, during microsurgery. The data were collected over twenty microsurgical sessions, ten were conventional and the remaining were RAM performed by four microsurgeons. The data were recorded in five bilateral muscle groups. The data was recorded for the evaluation of the localized muscle fatigue and the development of predictive models. From these data, a general dataset (508,328 records) was generated, for conventional and RAM. Two different preprocessing techniques (scaled and scaled and normalized) were applied. In this dataset, 80% of the data for training and 20% for testing purposes. This work advances the understanding of ergonomic risks in microsurgery by integrating wearable technology and predictive models of musculoskeletal risks. These results demonstrate the validity and accuracy of predictive models and constitute the starting point for achieving a comprehensive understanding of the ergonomic risks of microsurgery.