In this paper, using principal component analysis and cluster analysis method, combining with the statistics software MATLAB. The Qingyuan rough gentian, Zuojia rough gentian, Zuojia leaf gentiana system clustering analysis of experimental data. Main ingredients swertia glycosides, Swertia bitter glycosides, gentian bitter glycosides and oleanolic acid and ursolic acid, etc. Cluster analysis results showed that the three principal component contribution rate of the Qingyuan rough gentian is 89.24%, the three principal component contribution rate of the Zuojia rough gentian is 89.85%, the three principal component contribution rate of the Zuojia leaf gentian is 93.56%. By cluster analysis to Qingyuan rough gentian, Zuojia rough gentian, Zuojia leaf gentiana aristata 3 groups of data were divided into 6, 5, 4 classes, and determine the appearance characteristics of the high quality gentian, provide the basis for breeding to select high quality gentian.