The purpose of this study is to evaluate the predictive capability of a virtual test system (VTS) proposed for vehicle front assessment of pedestrian injury risk. The VTS accounts for a broad range of impact scenarios in pedestrian accidents and the assessment is done using recent pedestrian accident data. Firstly, simulation test samples (STS) accounting for the broad range of vehicle impact speed, pedestrian height and gait stance in real world impact scenarios were developed based on different sets of multibody vehicle‐to‐pedestrian impact simulations. Then a sedan and a van model were tested using the defined STSs. The AIS2+ injuries predicted from these STSs for each vehicle model were weighted by the involving proportion of each impact scenario observed from German In‐Depth Accident Study (GIDAS) pedestrian accident data via a defined Injury Weighting System (IWS). The injury predictive capability of the VTSs using different STS sample sizes and the corresponding IWS was evaluated by comparing the predicted AIS 2+ injury rate and distribution of AIS 2+ injuries as a function of pedestrian body region and height, vehicle class and impact speed with that observed from the GIDAS data. The results indicate that the proposed VTS using a STS of about 120 cases is broadly capable of predicting the AIS 2+ injury rate and distribution of pedestrian AIS 2+ injuries observed from the real‐world accidents when the same vehicle class distribution as the accident data is employed. The VTS can be considered as an effective approach for assessing pedestrian safety performance of vehicle front designs at the generalised level.