Fatigue behavior of two types of automotive steel, quenched and tempered SUJ2 and carburized SCM820PRH, which are applied as powertrain parts are studied. These two types of steel are different in their hardness distribution from surface to core. The hardness of quenched and tempered SUJ2 is homogeneous, in contrast to that of carburized SCM820PRH (SCM) which decreases from surface to core. These steels are investigated in terms of their monotonic tensile properties and fatigue behavior. A number of predictive methods were used to describe the fatigue behavior of these steels. A simple predictive method is based on approximation of S-N curve from ultimate tensile strength. The well known Murakami’s defect area method was also applied for the prediction of the high cycle fatigue strength. It was found that the classic estimation of fatigue behavior which considers 700 MPa as fatigue limit at 106 cycles for materials with ultimate strength of more than 1400 MPa is reasonably close to fatigue behavior of SCM steel. However, newer approach which estimates half of ultimate strength as fatigue limit at 106 cycles resulted in a close prediction of fatigue behavior of SUJ2 steel. Also, Murakami’s defect area method resulted in reasonable predictions of fatigue limit for both steels while being closer to experimental data for SUJ2 which has a uniform distribution of hardness.