Previous studies have shown that both excessive linear and rotational accelerations are the cause of head injuries. Although the head injury criterion has been beneficial as an indicator of head injury risk, it only considers linear acceleration, so there is a need to consider both types of motion in future safety standards. Advanced models of the head/brain complex have recently been developed to gain a better understanding of head injury biomechanics. While these models have been verified against laboratory experimental data, there is a lack of suitable real-world data available for validation. Hence, using two computer models of the head/brain, the objective of the current study was to reconstruct four real-world crashes with known head injury outcomes in a full-vehicle crash laboratory, simulate head/brain responses using kinematics obtained during these reconstructions, and to compare the results predicted by the models against the actual injuries sustained by the occupant. Cases where the occupant sustained no head injuries (AIS 0) and head injuries of severity AIS 4, AIS 5, and multiple head injuries were selected. Data collected from a 9-accelerometer skull were input into the Wayne State University Head Injury Model (WSUHIM) and the NHTSA Simulated Injury Monitor (SIMon). The results demonstrated that both models were able to predict varying injury severities consistent with the difference in AIS injury levels in the real-world cases. The WSUHIM predicted a slightly higher injury threshold than the SIMon, probably due to the finer mesh and different software used for the simulations, and could also determine regions of the brain which had been injured. With further validation, finite element models can be used to establish an injury criterion for each type of brain injury in the future.