This master’s thesis project investigated numerical modelling of polymer materials for crash simulations. The increase in polymer materials in vehicles calls for more complex material modelling in crash scenario. These materials present new challenges which require new solutions. This work had two areas of focus; material modelling of polymers and available databases of polymers. LS-DYNA was the programme used for the project, as it was already employed at Scania CV AB for crash simulation. The software offers many pre-defined material models and the most applicable models were compared. The models known as *MAT_024, *MAT_089 and *MAT_103 were investigated, *MAT_024 being the model implemented for most new materials. Furthermore, the strain rate dependency of the polymers were modelled with either a table of stress-strain curves for various strain rates or the theory of Cowper-Symonds. Firstly, the models were calibrated using a virtual tensile test aimed to replicate a physical tensile test. Force and deformation were studied for the virtual test, and then converted to true stress and true strain. This was then compared with experimental data. Calibration served as a helpful tool during set-up of the validation testing. Two types of validation tests were employed; an instrumented puncture test and a three point bending test. Virtual tests were performed, both aiming to replicate each physical test. No damage and breakage modelling were included in the scope of the project. Therefore, the force-deflection curve from the instrumented puncture test were only correlated with the virtual testing up until a point of damage. Lastly, the three point bending test was replicated with simulation. Here, the engineering stress-engineering strain curve results were compared between physical testing and simulation. The use of strain rate dependency models was recommended to be investigated further. Based on results from this study, a table of material curves was preferable for more linear polymers. The Cowper-Symonds methodology was best suited for less linear polymers with intersecting stress-strain curves. The results from the simulations presented a good correlation at smaller strain for *MAT_089. *MAT_024 and *MAT_103 presented little difference between the models. The correlation with the yield point was generally satisfactory. However, the models were not applicable for all polymers tested. The models worked well for more linear material behaviour, and worse for materials with strain softening like behaviour. *MAT_089 had a better applicability across all polymer materials, especially when the simulated yield point was calibrated close to that of the physical. From the studies performed a methodology for defining polymer materials for crash simulations could be determined. The methodology originated in *MAT_089, with the exception for more linear polymers where *MAT_024 was more favourable.