This work focuses on the high-cycle fatigue behavior of steel components through advanced numerical modelling, supported by experimental tests. The research is an essential part of national program SIRENES, which aims at improving the structural performance of floating platforms candidates for installation in deep-offshore environments. The SIRENES project extends the research performed in previous projects [1] on the development of a hybrid tension-leg platform (TLP), designed for the combined exploitation of offshore wind and wave energy, and addresses some important issues related to the structural integrity and lifetime strength of this hybrid platform. In the present paper, a rigorous and computationally efficient numerical methodology is presented for the reliable prediction of fatigue fracture initiation and its propagation in the steel component. The proposed numerical approach determines the crack tip location, its subsequent propagation direction, and the corresponding fatigue crack growth rate under high-cycle fatigue conditions. The numerical model incorporates an enhanced version of cohesive elements, capable of simulating accurately and efficiently the crack growth rate in the high-cycle fatigue regime, taking into account fatigue damage accumulation under cyclic loading conditions and using an efficient cycle-skipping technique [2]. An experimental testing program has also been conducted in the laboratory to validate the above numerical model. More specifically, fatigue crack growth rate tests using compact-tension (CT) specimens and crack tip opening displacement tests employing single-edge notch bend (SENB) specimens have been performed on specimens made of mild steel (S355) and high-strength steel (S700), and very good comparison with the numerical results were obtained.