Designing and modelling foundation structures crosses two engineering disciplines. The structural engineer who designs the structure and the geotechnical engineer who determines the bearing capacity of the soil. When modelling large shallow foundations, it is not always clear how to determine the stiffness of the soil and how it should be used in structural software programs. In general, the stiffness is determined by the geotechnical engineer and the determined value is used by the structural engineer. To determine the stiffness, it is important to know the intended use. For the use of the stiffness, it is also necessary to know what restrictions/assumptions are applicable. Besides uncertainties when determining the soil stiffness, other methods to model the interaction between structure and soil are also not straightforward. The goal of this thesis is to develop a consistent way of modelling large concrete slab foundations and optimize the interaction between structural and geotechnical engineers. The model should take into account the interaction between structure and soil and should be able to yield correct, fast (computational and modelling time) and useable results for the designer. The conclusions of this thesis can be split up into three parts. The first is related to the analytical analyses of different foundation models that were investigated, the second is related to the computational analyses and the third is related to the communication between structural and geotechnical engineer. Three simple foundation models with constant parameters have been analysed: A Winkler foundation, a Pasternak foundation and a Gradient foundation. The models were analysed analytically. In this study it is found that the results of Pasternak foundation model represents reality more consistently. This model only needs three input parameters and takes the spreading of the loads in the soil into account. This is done through the shear modulus of the shear layer (Gp value), which is difficult to determine. What can be concluded from the computational study is that, if the displacement is important, the Secant Method (SM) will calculate/provide values which are significantly smaller, compared to the Terzaghi and PLAXIS results, and another method must be used to verify the displacement. If the moment is important, the SM values can be used. The interaction between the structural and geotechnical engineer can be guided by a checklist. This list summarises the important aspects which need to be discussed to come to a quick and an appropriate model for a final design. This is an initial step towards improving and optimizing the work approach. What both engineers should keep in mind is that a Winkler model with non- uniform coefficients appears most suitable for most slab foundations.