Recently, a number of researches on detecting such inattentive driving have been carried out, and various detection methods have been proposed. In this paper, using driving data which can be measured with existing on-board vehicle sensors, the authors focus on headway distance behavior which relates to rear-end collision as well as the steering behavior which can be used to detect driver drowsiness. The emphasis of this study is that the inattentive driving can be potentially detected from the driving behavior deviated from the normative characteristics considering the transfer function from the environment to the driving operation. Therefore, intrusive measurement methods such as physiological measurement are not required. In addition, to detect the inattentive state by the criterion of the individual driver, the driver assistance system can be realized in the way that it can be adapted not only to various types of drivers, but also to the driver real-time driving state. In this study, the driver models are employed to estimate the normative driving. The driver models in car-following and in lane-keeping situations are constructed independently, and basically the constructed models are assumed to be the general dynamical system models. The driving data in the normative driving and in the inattentive driving were compared with these reference driver models. According to analysis results, this study exhibits the feasibility of the proposed method that it can detect the inattentive driving level by using the reference driver models.