Abstract: The threshold values are very important to define the indices of extreme climate events. Usually, the 90 percentile threshold value is used as the threshold values of extreme high temperature events, while the specific methods of categorization and calculation lead to different threshold values. In order to find the most suitable method, the quality controlled and homogenized daily maximum temperature data in China for the period of 1951—2008 are checked, and some missing data are reconstructed. Analysis is conducted using a new data for 1961—2008 in 224 selected stations. The preliminary analysis shows that the daily data series of most stations are not fit for normal distribution as generally believed. Based on the analysis above, the comparison for the threshold values falling into the percentiles of the extreme events of maximum temperature is accomplished using three different methods (including two traditional methods of calculating threshold values based on the assumed normal distribution and a new method by which the threshold values are calculated based on the sample probability distribution).The results show that the evaluation of the threshold values on the method using the real sample frequency distribution to simulate the probability distribution (the new method) has got the best results than using the other traditional threshold values calculating methods. Based on the comparisons, the threshold values of extreme events of maximum temperature in China have been developed with the 30-year sample climate phase. The following inspection shows the threshold values can reflect the real threshold values on the maximum temperature in daily use, and they are more stable and representative. In addition, the distribution charts of the threshold values developed by the different methods are drawn, and it's found that except for in Southwest and Northeast China, Qinghai, Inner Mongolia, the threshold values of extreme high temperature events for most part of China are all above 35 ℃, which proves that it is reasonable of using the threshold 35 ℃ as the standard of high temperature disasters.