Résumé : This paper aims to assess the uncertainty on the extreme responses of a semi-submersible wind turbine using two environmental contour approaches. The approaches are based on the inverse First Order Reliability Method and the direct Monte Carlo Simulation with different sample sizes. The long-term sea states are described by a 3-parameters Weibull distribution for the significant wave height and a conditional log-normal distribution for the zero-up crossing period. Two extreme sea states are extracted from each environmental contour, which are associated with the maximum significant wave height and maximum zero-upcrossing period. A 3-h time domain simulation for each sea state is conducted by adopting a validated numerical model. The root mean square for all the response spectra are estimated and compared to each other. It is observed that despite the small deviation between the significant wave heights resulting from adopting different environmental contour approaches, the deviation between the resulting responses is much more significant. The Gumbel and the generalized extreme value distributions are fitted to the maximum extreme response values for each sea state and unified fitted parameter are obtained.