This paper introduces a novel modeling and validation approach for hybrid electric vehicles (HEVs). The proposed dynamic modeling approach offers a more realistic simulation performance over most map-based studies of previous literature, while the novel validation approach requires no a priori information on the control algorithms running in the system but uses only measurement data collected from the actual system. A significant benefit of the proposed validation method is that it could further be used for the estimation of variables, which are unavailable for measurement, for variables such as engine torque, battery state of charge, generator torque, motor torque, fuel consumption, etc., as demonstrated in this paper. For the validation and estimation process, the simulation model must be driven with control signals obtained from the actual system, which, most of the time, are not available. To overcome this problem, in this paper, sliding-mode-control-based robust controllers are designed to emulate the engine, motor, and generator control signals to achieve minimum deviation between the variables that are calculated through the simulation model and measured from the actual system, in spite of the nonlinearities and uncertainties that are not considered in the developed model. This paper is based on the model and measurement data obtained from a series HEV, namely, the U.S. Military's High Mobility Multipurpose Wheeled Vehicle XM1124. The evaluation of the simulated and actual measurement data indicates the good performance of the developed modeling and validation technique, which is also motivating the use of the approach for the estimation of variables unavailable for measurement in a variety of systems.