On the basis of expounding the status quo of China’s orange drought warning model research, according to the real-time monitoring data of Beibei Orangery in Chongqing City from June to August in 2010, by using factor analysis and principal component regression, we construct the drought warning model of Beibei Orangery in Chongqing City, and we conduct test and correction on drought warning model by using MAPE principle of assessment and prediction accuracy and the real-time monitoring data of September. The results show that as for the prediction of orange drought, the model has strong credibility. If we conduct continuous dynamic monitoring on the relevant indices regarding orange drought by virtue of this model, and pay attention to the analysis of variation trend of prediction value, then it will be more practical; the warning effect, to some extent, further indicates that using factor analysis and principal component analysis method to conduct drought warning is rational; as the prediction model is static and linear, so we should perfect model to promote the accuracy and practicality of model.