Obesity is associated with an increased fatality risk in automobile collisions. The goal of this study was to evaluate the predictive capabilities of the obese human body models via comparison of model simulations to post‐mortem human subject tests from the literature. First, obese post‐mortem human subject sled tests were used to assess the ability of the obese GHBMC models to replicate occupant kinematics in a sled environment. Second, lap belt pull tests were used to evaluate belt interaction with the abdomen and pelvis. Third, adipose tissue‐level tests were used to evaluate the shear stiffness of the flesh material model used in the obese human body models. Similar to the post‐mortem human subject tests, the obese human body models experienced substantial lower extremity forward motion in the sled test simulations. However, the model did not exhibit submarining behaviour as observed in the post‐mortem human subject experiments. Further, the lap belt pull simulations failed to reproduce the belt/abdomen interaction seen in the post‐mortem human subject, and the material model used to represent the human body model flesh was found to be approximately one order of magnitude stiffer than human abdominal subcutaneous adipose tissue. This study shows that improved models of abdominal flesh, and specifically subcutaneous adipose tissue, may be required to obtain biofidelic belt/abdomen interactions and to predict submarining behaviour in crash simulations.