In the modern development process, the virtual component design plays a decisive role and a further increase in importance is expected. Hence, the improvement of methods describing the manufacturing process related fatigue strength is mandatory. Surface roughness plays a vital role on describing fatigue strength of various components. The aim of this work was to develop a methodology, which gives an enhanced understanding of the local surface roughness effect, based on measured three dimensional surfaces topologies. The input parameters of the developed method should necessitate only a small number of well-known material parameters to avoid costly material tests or complex additional measurements. The investigation starts with an improved method to get a triangulated surface from a measured, spatial point cloud. Such a measurement data contains high frequency signals, which have no influence on the fatigue strength, but makes the further processing of the surface data unnecessarily difficult. Thus, a digital Gauss filter was applied on the measurement data. To suppress the influence of the measured macroscopic surface waviness, the shape deviation is additionally corrected. The resulting surface data is integrated in a three dimensional linear elastic finite element model. Three standard load cases are set-up including two tensile one shear load step. The mesh size of the finite element model has an utmost impact on either the quality of the results and on the calculation time. Areas of comparably high stress levels also possess an increased surface curvature value. Therefore, a user-defined method is developed to determine the mesh size based on the maximum local curvature. To study the preferred roughness orientation and the effect of micronotches within the surface, a linear superposition of the three standard load cases is applied. This features also a multiaxial evaluation. This comprehensive methodology facilitates numerically based fatigue strength reduction factors for each surface node, which are based on the critical distance theory. The corresponding stress evaluation paths perpendicular to the surface are evaluated by using an advanced Superconvergent Patch Recovery approach. The critical distance length is based only on the fatigue strength and the crack propagation threshold of the defect free material. This meets the desired goal of a minimized number of material input parameters within this approach. The roughness effect under cyclic loading is locally assessable by these novel reduction factors for each surface node. Each of these local minima regions are bounded by certain threshold limit values to their adjacent regions. The relative occurrence of these local minima describes the influence as roughness fatigue factor and additionally enables a statistical size effect of the rough surface. The presented novel approach is numerically effective and therefore applicable for different surface conditions. A comparison of the developed roughness evaluation method against a rough surface of an aluminium casting concludes the work.