A non-adequate psychophysical condition represents a major factor in causing car accidents. In particular, 20% of car crashes are caused by mental fatigue and drowsiness, with dramatic consequences and fatalities. Nowadays strategies to reduce the risks while driving. The on-board systems equipping current vehicles are not able to intervene before the sudden onset of drowsy episodes and are affected by a poor accuracy resulting in several events’ misclassifications (i.e., false positive) causing drivers’ mistrust of technology.Being able to recognise in advance the occurrences of fatigue and drowsiness episodes would dramatically increase road safety and reduce car crashes. This is extremely relevant especially for professional drivers who drive for prolonged periods leading to an increase of risks due to not-proper psychophysical conditions. Even if professional drivers are trained to prevent, recognise, and minimize the effect of fatiguing, it must be considered that often the driver becomes conscious of drowsiness and mental fatigue onset too late, that is when it is already driving in a not-safe condition. The aim of this study was to adopt a multimodal approach to characterize the initial phases of fatigued mental state while driving, to develop an effective and timely detecting methodology. Ten volunteer professional drivers have been recruited to take part in an experimental protocol, performed in a car simulator. The experiment took place in the afternoon to increase the chance of eliciting mental fatigue and it consisted in driving for 45 minutes in a monotonous city-like environment. Before performing the monotonous driving task, participants were asked to drive for 15 minutes in a high-difficulty track race to induce fatigue, increasing the probability of a not-adequate psychophysical condition during the following monotonous driving task. Aiming at developing a neurophysiological model for mental fatigue characterization, a multimodal neurophysiological assessment was performed collecting the Electroencephalographic (EEG) and Electroculographic (EOG) signals. In parallel, behavioural assessment was performed through a secondary reaction task to detect eventual variation of performance from individual normal levels because of an altered psychophysical condition. Subjective measures were collected as well for the self-assessment of both fatigue and drowsiness state and task perception (high vs low demand). Behavioural and subjective measures have been so employed to (i) validate the experimental design; and to (ii) support and validate the employment of neurophysiological measures for characterizing the mental fatigue.