Accurate determination of contact stresses in the soil base is essential for the reliable design of raft foundations. Traditional analytical methods, such as the corner points method, offer simplicity but are based on simplified assumptions, which may lead to errors under complex engineering conditions. Numerical finite element modeling provides a more realistic representation of soil behavior but requires considerable resources. Comparing these approaches is relevant for defining their applicability limits and improving calculation accuracy. The article presents a comparison of stresses in a soil base under a rectangular raft foundation obtained by the analytical corner points method and finite element numerical modeling for elastic–linear and elastic–plastic soil model with the Mohr–Coulomb failure criterion. The aim of the study is to refine the influence coefficients used in the corner points method, to identify quantitative and qualitative differences between the approaches, and to outline the boundaries of their correct application. The comparison revealed significant discrepancies not only in the values of contact stresses but also in the shape of their distribution surface. Preliminary discrepancy coefficients for the studied case have been determined. The results confirm that even for relatively simple geometry and loading conditions, discrepancies between the methods can have a substantial impact on the assessment of bearing capacity and the uniformity of soil foundation performance. The proposed approach lays the groundwork for further comprehensive studies of the influence of geometric parameters of the foundation, type and magnitude of loading, and the physical–mechanical properties of the soil on the consistency of results obtained from analytical and numerical calculation methods.