Abstract: Sampling uncertainty of hydrological variables parameter commonly exists in frequency analysis-based hydrological drought indices calculation due to the limited data length and discontinuousness. Based on the daily streamflow data of the Yichang Gauging Station from 1882 to 2009, the runoff drought index SDI of different time scales was calculated, and the sampling uncertainty of drought index calculation was verified by Monte-Carlo stochastic simulation, analyzing the influence of drought evolution characteristics and uncertainty on drought assessment in the upper reaches of Yangtze River. The results show that when evaluating drought magnitude according to drought index, considering the influence of uncertainty can provide theoretical support for more accurate and effective on drought resistance decision, to avoid the possible lack capacity of disaster reduction or resources waste of drought-resistant.