In recent decades, frequent droughts have been detected in some places in China along with the global climate change, which cause great damage to the development of social economy. Thus, it's high time to investigate the spatio-temporal evolution of droughts. Based on the newly published scPDSI PM (self-calibrating Palmer drought severity index based on Penman-Monteith) dataset with high spatial resolution, the annual and seasonal variations of droughts in China from 1961 to 2009 were researched. Other research methods like linear regression, Mann-Kendall method, wavelet analysis as well as REOF (rotated empirical orthogonal function) were adopted to investigate the spatio-temporal pattern of droughts in China, and find out how droughts changed under the influence of the general circulation of atmosphere, in order to provide the foundation for drought control and drought resistance. The results indicated that: 1) In 1961-2009, China as a whole became moist significantly at annual and seasonal scale with the mutations of turning drought to wetness detected in early 1970s. 2) Period analysis illustrated that the oscillation period of seasonal and annual drought remained generally unanimous, and the dominant period of annual drought was 6.2 years, while drought in spring, autumn and winter generally had the periods of 2.6 and 6.2 years with 6.2 years as the dominant period; otherwise, drought in summer presented a dominant period of 4.4 years. 3) In the light of the spatial modes disassembled from REOF, the entire country was divided into 8 regions with different characteristics of drought or wetness, which were Northwest China, Northeast - Inner Mongolia Plateau area, the Greater Khingan Range area, the northern Qinghai-Tibet Plateau area, the southern Qinghai-Tibet Plateau area, the Middle China (including Sichuan Basin, Hanzhong Basin and Loess Plateau), Huang-Huai-Hai Plain and Southeast China. It was found that the sub-regions of REOF could primely manifest the geographical features of different regions, and hence objectively testify the actual drought condition in China. Among the 8 regions, the Greater Khingan Range area, Southeast China, the southern Qinghai-Tibet Plateau area, and Northwest China were getting moist while the latter 2 areas had a significant trend (P<0.05); Northeast - Inner Mongolia Plateau area, the northern Qinghai-Tibet Plateau area, the Middle China and Huang-Huai-Hai Plain were getting dry, and except Huang-Huai-Hai Plain, the former 3 regions were prominently getting dry. Take the regions as the objects, no matter at annual or seasonal scale, Northwest China was getting moist, while the Middle China was getting arid (P<0.05). Although the geographical positions of the 8 regions resulted in the discrepancy of influence under different climate factors, the 8 regions generally had the oscillation period of 2-9 years. 4) M-K test (Mann-Kendall test) on the 8 regions showed that the seasonal MK values of different regions varied with each other, which indicated that drought-wet degree of different regions had obvious seasonal feature, and yet most regions had a tendency of getting drought in spring and getting moist in summer. 5) Good correlations were found between polar vortex index and drought in most areas. Correlation analysis indicated that droughts in Northwest China may be affected by the polar vortex, Indian Ocean Dipole (IOD) and Pacific Oscillation (PDO). Droughts in Northeast - Inner Mongolia Plateau area, the Qinghai-Tibet Plateau area and the Middle China were significantly influenced by IOD, and drought of Huang-Huai-Hai Plain was negatively correlated with PDO. Unlike the conventional PDSI index, scPDSI integrates solar radiation and air speed into account, which will be of scientific and practical importance to forecast and distinguish different drought conditions.