To address the shortage of data regarding extreme winds in P. R. China, we examined Xiamen, located on the southwestern Chinese coast, as an example. We used the Batts wind field model and the MonteCarlo method to simulate extreme wind speed distribution models and to predict maximum wind speeds in different recurring periods. The simulated results coincided with the experical distribution function. Comparing the numerical analysis results of different extreme distributions, such as the Gumbel distribution function, the Frechet distribution function, the reverse Weibull distribution function, and the Generalized Parato distribution function, shows that the reverse Weibull distribution function provides the most precise prediction of maximum wind speed for a onehundredyear return period, while the Frechet distribution function significantly differs from the results of all the other distribution functions. The extreme wind speeds of different return periods are estimated.