Objective: Finite element analysis (FEA) has been frequently used to study the loading situation of dental implants and bone resulting from the fixation of nonpassively fitting restorations. The goal of the present investigation was to demonstrate the effect of geometric model parameters and mesh size on FEA results. Method and Materials: Five three-dimensional FEA models representing a three-unit fixed dental prosthesis (FDP) supported by two terminal implants were constructed. The models differed in terms of mesh size, bone geometry, implants, and restoration and were created either by joining virtual free-form objects or utilizing optical scans of existing components. By applying thermal changes in volume of specific elements in the area of the FDP pontic, a horizontal misfit of 10 µm between implants and the restoration was introduced. The resulting loading situation of the bone around the implants was recorded as von Mises equivalent stress. Results: Maximum stress magnitudes ranging from 13.1 to 24.9 MPa occurred in the cortical part of the implant site where the neck of the implant penetrates bone. In trabecular bone, loading magnitudes were lower by a factor of 20. Modeling implant threads did have a remarkable effect on the stress situation as well as different span lengths of the restorations modeled. All other parameters led only to small variations in maximum loading magnitudes. Conclusion: Simplistic FEA models based on virtual free-form objects with limited level of mesh refinement seem to allow for a basic evaluation of peri-implant bone loading resulting from the fixation of misfitting superstructures. Keywords: bone loading, finite element analysis, implant-supported restoration, model geometry, passive fit, von Mises equivalent stress