The characteristics of drought in the Weihe River basin are analyzed with Markov chain and stochastic process theory. The results show that the continuous wet (dry) periods are usually extended from 2 to 3 years. The probability of a single wet (dry) year is larger than that of continuous wet (dry) years. The probability of continuous dry years is larger than that of continuous wet years. The periods of continuous dry years are longer than that of continuous wet years. The sudden change from a dry year (wet year) to wetter year (drier year) has negative effects for irrigation. The probability of drier year is the greatest, and then wetter year, the probability of wet year is the smallest. The frequencies of the continuous 2- or 3-dry year are high and their intensities are great. In addition, with the partial correlations analysis, the relationship between the precipitation index, the temperature index and the runoff in the Weihe River basin are analyzed. On the basis of these data, the drought index is estimated for the area.