Under the global climate change, quantitative research on response characteristics of precipitation to global climate change and its future trend in northwest China is very important. Using empirical orthogonal function (EOF), morlet wavelet analysis, the run theory, displacement theory, extremum deviation theory, seasonal index method (RI), the Mann-Kendall rank correlation method and climatic trend coefficient method, the spatial distribution and the multiple-time scale variability of precipitation are analyzed by the monthly precipitation data from 95 meteorological stations in northwest China during 1951 2008. Results indicate that: (1) The contribution of the overall changes rates of the first four characteristic vector was 81. 61%, the first vector of precipitation spatial distribution pattern was primarily complex (contribution to the overall changes was 48. 53%), the research area can be divided into five sub-regions accordingly. (2) There were primary cycles of 9 years and 12 years for the precipitation series. (3)The probability of single abundant precipitation year was higher than single short precipitation year and the continuous short precipitation year was higher than continuous abundant precipitation year. Drought occurred mainly in Xinjiang and the Hexi Corridor. The Hurst coefficient was 0. 6290. 845, so the precipitation series had durative characters. (4) There was 40% area of northwest China with a clear seasonal distribution; 61% area reproducibility index (RI) value were greater than 0. 7,showing that in most area annual rainfall distribution was stable and had a less differences for inter-annual precipitation. (5) The 86. 3% area showed an increasing trend (20% and 48. 4% of the stations passed 0. 01 and 0. 05 significance test), and increase ratio was 16. 6 0.1 mm/lOa, and 11. 6% area decreased with 5. 20. 3 mm/lOa.