In this study, we sought to augment existing driver fatigue detection techniques, which are deficient in individual fatigue feature analysis, precision, and resilience. A novel driver fatigue monitoring intelligent system is introduced, employing human–machine hybrid enhancement. To address these limitations, a human–machine hybrid fatigue driving experimentation platform was designed using a hardware-in-the-loop system. This amalgamation delivers accurate fatigue level assessments. Subsequently, three preprocessing methods were compared for facial imaging and vehicular status data, developing a driver human–machine hybrid fatigue driving database. This comprehensive database includes facial images, steering wheel angle, and acceleration pedal data, aiding in detecting fatigue-induced behavioral shifts. Lastly, variance analysis was employed to quantify the significant difference levels of human–machine hybrid fatigue feature parameters across varying fatigue levels. On the basis of this analysis, a machine-learning-technique-based human–machine hybrid enhanced driver fatigue monitoring intelligent system was developed, achieving accuracies of 95.5%, 91.5%, 94.7%, and 95.0% in distinguishing four driver fatigue stages, namely, wakefulness, mild fatigue, fatigue, and severe fatigue, respectively. Our findings validate the efficacy of our proposed system in discerning driver fatigue levels and its potential to significantly improve transport system safety and efficiency.