The standardized precipitation index (SPI) was used to quantify the classification of drought in the Guanzhong Plain, China. The autoregressive integrated moving average (ARIMA) models were developed to fit and forecast the SPI series. Most of the selected ARIMA models are seasonal models (SARIMA). The forecast results show that the forecasting power of the ARIMA models increases with the increase of the time scales, and the ARIMA models are more powerful in short-term forecasting. Further study was made on the correlation coefficient between the actual SPIs and the predicted ones for the forecasting. It is shown that the ARIMA models can be used to forecast 1-month leading values of all SPI series, and 6-month leading values for SPI with time scales of 9, 12 and 24 months. Our study shows that the ARIMA models developed in the Guanzhong Plain can be effectively used in drought forecasting.