In this thesis, the impact behaviour of 3D-printed PMMA (poly-methyl-methacrylate) specimens is predicted using Finite Element modelling. As an example, cranial implants have to withstand impact loads, while the 3D-printing process allows production of complex geometries. The ideal infill geometry to perform well under impact (i.e. absorb more energy than other geometries) is not known beforehand. The goal of this thesis is to find the best performing infill structure using Finite Element modelling instead of manufacturing and testing possible infill geometries. In collaboration with the chairs of Polymer Processing as well as Materials Science and Testing of Polymers of Montanuniversitaet Leoben, solid specimens were 3D-printed and tested according to the standard for impact testing of polymers. The experimental setup was modelled and the results of the solid specimens were the basis for calibrating two material models to predict the absorbed energy during impact. The test specimens' damage and failure behaviour is reproduced using a material damage model. The brittle damage model in ABAQUS represented the test results best. To use the brittle damage model, the material law has to be isotropic and linear elastic. Nevertheless, the damaged material behaviour is anisotropic because the distributed damage is modelled via stiffness degradation in the direction of loading. To evaluate these material models, an alternative infill structure was manufactured and tested. The test results were then compared to the predicted simulation results. Furthermore, the results of both infill structures were considered for the calibration of a third material model. To evaluate the prediction quality of those three material models, two new geometries were manufactured and their simulation results compared to the test results. Two out of the three calibrated material models gave a qualitatively correct prediction for the absorbed energy. This means that the FE models answer the question which infill geometries perform better (i.e. absorb more energy) during impact. However, when it comes to the absolute values of absorbed energy, the simulation results deviate from the test results. For the main issue of this work, which was to find infill geometries which perform better than others during impact, the qualitatively correct results are a satisfying outcome. The quantitative results might be improved in future work by implementing an enhanced material law in combination with the calibrated damage models.