In this paper, we combine two feature extraction algorithms, Empirical Mode Decomposition (EMD) and wavelet packet transform, to analyze the EGG signals of students exposed to vocal music. We extract features from these signals by determining the instantaneous frequency and node signals based on the mean value of the envelope. The EGG signals were cut into short-time smooth signals. The average sample entropy value of the processed EGG signals was calculated to reflect the EEG activities of students under vocal stimulation. Then the EGG signals were used to reflect the changes in the mental health status of college students. In the vocal stimulation experiment, it was found that the students’ psychological comprehensive relaxation reached 85.47% on average,