People pay attention to the relationship between drivers' state before driving and subsequent vehicle operation as drivers' state like insomnia and overwork may contribute to traffic accidents. In this study, we classified drivers' state based on their sleeping hours before each driving session, and then we tried to bring out the effects in vehicle operation data, 20 in total, due to drivers' sleeping hours. We used the hierarchical clustering by Nearest Neighbor method to classify sleeping hours before driving. We classified the length of sleeping hours in each subject under three states, and then we calculated the mean and the standard deviation of each vehicle operation feature per subject. As a result, there was a common trend for all subjects (in DBP, distance from a point where they stepped on the brake pedal to the point where the vehicle paused temporarily). Surprisingly, we found that lack of sleeping hours make the drivers step on the brake earlier than when they had adequate sleep (because the mean difference DBP between the two states was 5.6 meters).