Two experiments were conducted to study driver fatigue. The first investigated driver fatigue and driving performance. Thirty one Participants completed a questionnaire to obtain their Subjective Driver Fatigue Score (SDFS) quantifying fatigue levels. Driving performance was evaluated by measuring steering wheel, lateral position, etc. The results showed significant increases in the SDFS and driving performance impairment following simulated driving sessions. Further analysis suggested a linear relationship between the SDFS and the standard deviation of lateral acceleration. Subjective fatigue assessment and driving performance were plotted as radar diagrams to show the multidimensional characteristics. The second experiment examined effects of caffeine and music on the SDFS, driving performance, and 8 EEG signal parameters. Initially, there was no significant inter-sessional variation in the dependent variables, suggesting all sessions were started at similar states. The final SDFS for caffeine and music sessions were significantly lower than control sessions, suggesting both inhibited subjective fatigue increase. Driving performance deteriorated less significantly in caffeine sessions than in control and music sessions. The results suggested that caffeine was more effective than music. EEG was not changed significantly. However, the amplitude of α wave increased significantly for an extremely fatigued individual, along with vehicle drifting and micro-sleep. In conclusion, the SDFS developed in this study successfully estimated subjective driver fatigue levels and showed a linear relationship with driving performance during driving tasks. Caffeine and music reduced driver fatigue subjectively similarly, but caffeine also helped subjects maintain driving performance.