In the context of conducting a side-impact collision test without side airbag for a specific vehicle model, an analysis of the injury value curve of the crash test dummy, when bench-marked against results from whole-vehicle testing, sled test and sled simulation, reveals two issues with the sled results: a delay in contact time and an excessively high peak value. This paper aims to analyze the problems arising from the simplification of the whole-vehicle model to the side-impact sled model. For the first time, these issues are addressed by correcting the sensor waveforms of the whole-vehicle test vehicle. Subsequently, by adjusting the input curves of the sled simulation model based on benchmarking the injury values of the sideimpact sled dummy, a more accurate simulation of the dummy injury value curve is achieved compared to the experimental results. This approach represents an improvement in the substructure method for sled simulation, and its practicality and convenience have been validated through its application in a specific development vehicle project. The total score of the dummy injury in the experiment is 15.66, and the simulation is 15.76, the results are very close. The contact timing and peak value of chest, abdomen, and pelvis injury in the simulation are basically consistent with the experimental results. It offers a novel method for benchmarking dummy injuries, sled simulation, and sled testing in the context of side-impact collision condition, providing a new avenue for research and development in the automotive industry. INDEX TERMS Vehicle side impact test; side sled test; side sled simulation; sensor waveform; dummy injury curve;