Stress is an inevitable problem experienced by people worldwide. Continuous exposure to stress can greatly impact mental activity as well as physical health thereby leading to several diseases. In this study, we investigate the effectiveness of audio binaural beat stimulation (BBs) in mitigating mental stress. We developed an experimental protocol to induce four mental states: rest, control, stress, and stress mitigation. The stress was induced by utilizing Stroop Color Word Test (SCWT) with time constraints and mitigated, by listening to 16 Hz of BBs. The four mental states were assessed using behavioral responses (accuracy of target detection), a perceived stress state questionnaire (PSS-10), and electroencephalography (EEG). The mean spectral power of four frequency bands was estimated using Power Spectral Density (PSD), and five different machine learning classifiers were used to classify the four mental states. Our results show that SCWT reduced the detection accuracy by 59.58% while listening to 16-Hz BBs significantly increased the accuracy of detection by 27.08%, (p = .00392). Furthermore, the support vector machine (SVM) significantly outperformed other classifiers achieving the highest accuracy of 82.5 ± 2.0 % using the beta band information. Similarly, the PSD topographical maps showed different patterns between the four mental states, where the temporal region's PSD was mostly affected by stress. Nevertheless, under mitigation, there was a noticeable restoration in the temporal activity. Overall, our results demonstrate that BBs at 16 Hz can be used to mitigate stress levels.