Abstract: Extreme sea conditions pose a significant threat to the safety of offshore structures. Generally, the extreme sea conditions under the case of future climate change and within a certain time range are predicted by extreme value theory in order to evaluate the response of the offshore structures to the extreme sea winds, currents and waves. However, the values of extreme recurrence period obtained through the extreme value analysis may vary due to the differences in extreme value samples, probability distribution functions and parameter estimation methods, and the differences in extreme values may further affect the safety and economy of the structures. For this reason, the hydrological simulation time series data spanning 43 years from 1979 to 2021 are collected from a proposed wind turbine site located in the Wenzhou Offshore Area of Zhejiang Province and used to study the sensitivity of once-in-50-year wave height to different extreme value samples, probability distribution functions and parameter estimation methods. The results indicate that under the cases of good fitting performance, both the sampling method and the probability distribution function have a significant impact on the once-in-50-year wave height, while the parameter estimation method shows a relatively smaller impact. To ensure the safety margin of the structures, it is recommended that the sensitivity analysis of the extreme wave height should be prioritized. In this case, the Pearson III distribution and the annual sampling method show a high stability and a low sensitivity in the extreme value estimation. The results of this study can provide a reliable reference for determining the extreme recurrence period reasonably.