Abstract: Based on the instrumental temperature data from 160 stations in China mainland and 9 stations in the Mongolia Cold Wave (MCW) key area, this papaer analyzed the regionalization of winter temperature change in China mainland during 1951-2005, using the statistical methods including cluster analysis, correlate analysis and Mann-Kendall test. It is found that China mainland could be divided into 5 regions consisting of 11 subregions and the spatial pattern was featured by consisting with spatial pattern of macro landforms over China. And 10 of the 11 subregions, excepting for the Sichuan-Guizhou subregion, had statistically significant warming trend during 1951-2005. From the north to the south, the abrupt warming year started gradually later. The North abruptly warmed since the 1970s, while the middle since the 1980s and the south since the 1990s. Besides, the warming rates were gradually slower from the north to the south with the maximum of 0.53℃/10yr in the north and minimum of 0.15℃/10yr in the south. In addition, 146 of 160 stations in China mainland were significantly positively correlated with the temperature change of the MCW key area. The spatial distribution of the correlation coefficients is consistent with the spatial pattern of regional differences of temperate change. This consistency suggests that regional differences of temperate change are strongly impacted by the spatial heterogeneity of MCW’s effects on regional temperature through the interaction of both the winter monsoon circulation and the macro surface landform. They are dominant factors inducing regional differences of winter temperature change in China mainland.