Problem formulation: The aim of the present PhD thesis is the investigation of the fatigue behaviour of flat samples made from thermo-mechanical rolled fine grained structural steel S500MC, subjected to tension-compression loading sequences containing overload blocks. Next using the experimental data, an improved fatigue life prediction method for the practical engineering use, taking in account the effect of the overload sequences has to be developed. The enhanced method should be applicable to broad spectra of metallic materials. Findings: Firstly the influence of four different post-cutting treatments on the fatigue strength of the specimens made of S500MC has been investigated. The results have shown that post-cutting treatments essentially increase the fatigue life of the samples. The derived fatigue lives were about three times higher than the samples without post-cutting treatments. The most important insight from this investigation is that even an economically advantageous post-cutting treatment can lead to an improved fatigue life of the sheet structures. Next the influence of overload blocks on the fatigue strengths has been investigated. Several load sequences with various load factors have been evaluated. To identify the crack initiation phase, an optical mess system has been installed. The experiments with one-thousand overload blocks have only shown a clear influence of the fatigue strengths of the samples; however the crack initiation was caused due to the high pulsating quasistatic overloads. Hence, the fatigue strength of the samples can be increased using different post-cutting treatments. In the third and subsequently last step of the present work an improved fatigue life prediction method taking into account the overload blocks with a minimal level of effort for the practical engineering use at design stress level has been developed. The nonlinear behaviour of the S/N-curve has been taken into account. Similarities and differences in comparison with the linear cumulative damage rule were briefly examined and discussed. The discrepancy between the Pålmgren-Miner rule and the proposed novel method, especially in the high load ranges consists in the fact that the linear accumulation rule assumes a perfect linear S/N-curve in the concerned stress range. The validation of the proposed method has been done in the finite life range using uniaxial experimental fatigue data, carried out on various specimens, made of different isotropic metallic materials. The calculated results have shown that the predicted fatigue lives using the presented method are generally more accurate than the fatigue lives calculated using the Pålmgren-Miner rule; however some of the results are non conservative in some cases. An improved accuracy regarding the predicted fatigue life could be achieved using nonlinear damage growth curves, which simultaneously implies the implementation of additional material parameters in the calculation.