Abstract: The extreme value estimation for the non-Gaussian wind pressures is extremely important for the wind-resistant design of building envelope. With its simplicity and unrestricted application region, the moment-based piecewise HPM (PHPM), Johnson transformation model (JTM) and piecewise JTM (PJTM) are usually used to estimate the non-Gaussian wind pressure extremes. Currently, the systematic research on the performance of the PJTM in the non-Gaussian wind pressure extreme estimation is less addressed, and the differences in extreme values estimated by the three unrestricted application region models are unclear. To compare the differences of the three models and provide certain selection principles, this paper systematically compares the accuracy of the three models to estimate the non-Gaussian wind pressure extremes. Firstly, this paper compares the parent probability distribution functions (PDFs) and translation functions by the three models theoretically. Secondly, the very long wind pressure data from a wind tunnel test are used to evaluate the accuracy of the three models to estimate the extreme values of the non-Gaussian wind pressures. The results show that, for the minimum value estimation of the non-Gaussian wind pressure with negative skewness, the accuracy of PHPM is generally higher than that of JTM and PJTM, while for the maximum value estimation of the non-Gaussian wind pressure with negative skewness, the accuracy of PJTM and PHPM is generally higher than that of JTM.