—Currently traffic management is becoming more important to achieve the goal of sustainable transport, and a good traffic model can describe the traffic behavior efficiently. The traffic models can be classified based on level of details as submicroscopic-, microscopic-, mesoscopic-, and macroscopic-level models. In this paper, we provide a review of the four types of models (submicro, micro, meso, and macro) and then propose a multilevel model of traffic, which combines submicroscopic, microscopic , and macroscopic levels of traffic model. In this work, we do not consider the mesoscopic-level model. At the submicroscopic level, we develop a bond graph model of a four-wheeled vehicle considering the longitudinal, lateral, yaw, and actuator dynamics. At the microscopic level, we develop a car-following model based on virtual interconnections between the submicroscopic bond graph models of vehicles. Then, at the macroscopic level, we deduce macroscopic variables (average speed, density, and flow) from the submicroscopic and microscopic models. Having a multilevel model of traffic allows combining two properties of modeling simulation, one in real-time mode at microscopic and submicroscopic levels and the other at offline mode at macroscopic level. Thus, the whole supervision of the road traffic can be performed. Finally, the multilevel model of traffic is validated on a real-time simulator of vehicle dynamics, based on experimental measurements acquired from intelligent autonomous vehicles (IAVs). In addition, real experiments on IAVs are performed to validate the model.