This paper aims to evaluate the biofidelity of a human body FE model with abdominal obesity in terms of submarining behavior prediction, during a frontal crash event. In our previous study, a subject-specific FE model scaled from the 50th percentile Global Human Body Model Consortium (GHBMC) human model to the average physique of three female post mortem human subjects (PMHSs) with abdominal obesity was developed and tested its biofidelity under lap belt loading conditions ().In this study frontal crash sled simulations of the scaled human model have been performed, and the biofidelity of the model has been evaluated. Crash conditions were given from the previous study (), and included five low-speed and three high-speed sled tests with and without anti-submarining device.The biofidelity of the morphed human FE model in terms of submarining behavior was evaluated by the correlation on overall body and belt-to-pelvis kinematics between simulation and tests, and quantified by correlation and analysis (CORA) rating scores. The CORA ratings showed that the overall responses of the human FE model simulations were well-correlated with those from tests. But the belt-to-pelvis interaction that determines the submarining behavior was reasonably-correlated with that from tests since inter-subject variance in local-area responses was quite large.This study warrants further investigation on modeling parameters of pelvic bone and surrounding soft tissue to better predict the performance of anti-submarining safety device for subjects with abdominal obesity.