Severe pitting corrosion has been responsible for many failures of carbon steel pipelines, especially where protective coating systems have become dysfunctional and where cathodic protection is not provided. High stress concentration at pit zone accompanied by cyclic loadings such as cyclic change in internal pressure, temperature or both, can cause initiation and propagation of fatigue cracks leading to failure of pipeline. Attempts have been made to employ the Finite Element (FE) method to calculate the stress concentration factors (SCFs) that are key parameters in the assessment of fatigue life of structures, to assess the fatigue life of pitting corroded pipes. However, the three-dimensional form of the pitting corrosion has been modeled in a highly simplified manner, typically as semi-ellipsoidal or cuboidal volumes. Real pits can have very complex three-dimensional forms, for example with many small pits at the base of a bigger complex pit. The roughness of the base of these more complex but realistic pit geometries can change the stress concentration at the base of the pit significantly and produce considerably different stress concentrations. Idealization of a complex pit geometry into a simplified semi-ellipsoidal or cuboidal pit may result in significant inaccuracies in estimating the fatigue life of a corroded pipe. The present study addresses this issue. In the present paper, several external pits with different aspect ratios and real complex forms are modeled on X65 steel pipes using advanced finite element methods. For comparison, the same FE models with the same size of pits but with idealized semi-ellipsoidal and cuboidal geometries are modeled and the results are compared. The Adaptive Meshing algorithm is utilized to have the best accuracy for comparison of the results with the minimum run time. Most importantly, full-scale tests data are used to validate the finite element models of the present study.