The Standardized Precipitation Evapotranspiration Index (SPEI) has been increasingly proposed for detecting and monitoring drought under the global climate, while its sensitivity to corresponding meteorological factors remains unclear. A global sensitivity analysis based on SPEI drought index was conducted from a data set of daily climate variables in 5 meteorological stations in Hulunbeier. The reference crop evapotranspiration was calculated by the FAO Penman-Monteith equation, and the SPEI sensitivity coefficients to key meteorological variables were determined by using Sobol's global analysis method based on monthly scale SPEI. The results indicated that the average drought intensity was-1.19, and showed an insignificant decrease trend at the rate of-0.03·10a-1. The cumulative of extreme drought intensity ranged from-7.33 to-0.80, and the descending rate was-0.12 per 10 years. The drought frequency fluctuated from 8% to 58% with the mean value of 33%. The extreme drought events had significantly increased immediately after entering the twenty-first Century, although an insignificant increase was detected in drought events for the last 56 years. Sobol's sensitivity analysis showed that among the six meteorological factors, precipitation was the most sensitive one to the SPEI across the four seasons consistently, whereas the other most sensitive factors followed by precipitation were mean temperature in spring, relative humidity in summer, wind speed in autumn and mean temperature in winter, respectively. According to the total sensitivity coefficients of the annual SPEI to precipitation, mean temperature, relative humidity, wind speed, temperature difference and net surface radiation were: 0.71, 0.13, 0.12, 0.11, 0.03 and 0.01. The annual SPEI had strong correlations with temperature difference, relative humidity, precipitation by-0.671 (p<0.01), 0.551 (p<0.01) and 0.872 (p<0.01), respectively. The correlation coefficients of average temperature, net surface radiation and wind speed of SPEI were-0.151, 0.086 and-0.150, respectively, showing weak correlation. Through the sensitivity and correlation analysis of SPEI and meteorological factors, it was concluded that interannual SPEI was more affected by precipitation and relative humidity, and was less affected by net surface radiation in Hulunbeier.