Fracture site motion is thought to play an important role in the healing of complex fractures of the distal femur via mechanotransduction. Measuring this motion in vivo is challenging, and this has led researchers to turn to finite element modeling approaches to gain insights into the mechanical environment at the fracture site. Developing a systematic understanding of the effect of different model choices for distal femur fractures may allow more accurate prediction of fracture site motion from these types of simulations. In this study, we aim to assess the effect of four different modeling choices and parameters. We looked at the effect of using bone specific density distributions vs generic values, employing landmark‐based geometry generation, varying fracture alignment within clinically relevant ranges, and determining whether direct apposition of the fracture to the plate was achieved. For validation, five cadaveric femurs had fractures created and repaired with plated constructs, and these were then loaded and fracture site motion was directly measured. We found that using landmark based bone geometry and patient‐specific bone density distributions had a minimal effect on the overall model predictions. Changing the alignment, particularly into varus and procurvatum could have a large (>50%) effect on predicted shear motion, as could direct apposition of the bone to the plate. These findings demonstrate that modeling choices can play an important role in simulating distal femur fracture mechanics, and it is particularly critical that patient customized models attempt to accurately represent alignment of the bone fragments and lateral plate apposition.