<p>With the development of methods for predicting extreme hydrological elements using probabilistic approaches, several commonly used methods have emerged for analyzing the risk of storm surge disasters, including the Annual Maxima method, the Peak-Over-Threshold method, the Gumbel distribution, and the Weibull distribution. Meanwhile, and emphases have been placed on assessing and comparing the applicability and stability of these various methods. To evaluate the rationality of different methods, we an entropy uncertainty analysis method was introduced based on information utilization efficiency, in which the sample Stochastic uncertainty is measured by the ratio of information entropy before and after sampling, i.e., the information extraction efficiency of the sampling method. Additionally, the cognitive uncertainty of the research method is assessed by the ratio of mutual information between the model and the sample to the information entropy of the sample, i.e., the information extraction efficiency of the mathematical model. Furthermore, we incorporated the group probability calculation method, information entropy and mutual information theory to analyze and calculate the entropy uncertainty more accurately. By applying this analysis to the design wave height and the recurrence period projected in the sea area west Guangdong of China, we believed that the most reasonable hazard assessment method shall be based on the over-threshold method combined with the Pareto distribution. Conversely, the assessment method based on the process extreme value method is deemed insufficiently reasonable and requires further research.</p>