Résumé : This paper explores the role of wave spectral characteristics and wave time history on the estimation of extreme mooring loads on floating offshore wind turbines. This research is applied to the DeepCwind semi-submersible platform located at the BiMEP test site in the North of Spain. Extreme sea states are selected using the inverse first-order reliability method (I-FORM). Mooring loads are modelled by quasi-static and dynamic numerical approaches. Different wave time series are generated numerically for each sea state to investigate the variability in predicted peak loads. All cases simulated incorporate the combined effect of wind and waves. Differences of approximately 30% in peak loads are found for the mooring system, reaching 40–79% for the most extreme sea states. Safety factors are proposed to account for sensitivity to wave groupiness in modelling loads under extreme work conditions of the DeepCwind platform (e.g., pitch and velocity control). A comparison between theoretical and real-sea wave spectra is also modelled to investigate possible differences due to the presence of multiple spectral peaks associated with swell and wind seas. In general, results show differences below 12% in the prediction of loads between both assumptions.