Performance anxiety is a common issue among musicians, and it could be fundamental to monitor their psychophysiological states during performances through non-invasive methods to support them in managing anxiety. Hence, infrared thermography (IRT) could be a valuable tool for this purpose. The study aims to assess whether IRT can effectively monitor musicians' psychophysiological states. The facial temperature of four musicians was recorded during two conditions: rehearsal and live performance. The temperature time course was extracted from 3 regions of interest (ROIS) (i.e., forehead, nose tip, and perioral) and the following metrics were computed: skewness, kurtosis, and sample entropy. Moreover, machine learning models were applied to evaluate the presence of stress and the balance between sympathetic and parasympathetic systems. The results showed notable changes in thermal metrics in all the ROIs. Moreover, the prevalence of the sympathetic system for 50% of the rehearsal and 92% of the live performance durations was assessed. Additionally, the presence of elevated stress indicators was assessed for 6% of the duration of the rehearsals and 9% for the live performances. These results demonstrated the capability of IRT to assess modifications of the psychophysiological state of the musicians secondary to the condition of the performance.