Abstract: In order to effectively evaluate cigarette quality and its consistence, the physical indexes and mainstream smoke indexes of cigarette samples A, B₁ and B₂ were analyzed with Principal Component Analysis (PCA), and their principal components were determined. According to characteristic matrix, the principal component score diagrams of cigarette samples were established, and cluster analyses were conducted. By using principal component load graphs, the importance of cigarette physical indexes and mainstream smoke indexes and their correlations were analyzed respectively. The results showed that:1) The samples of different cigarette brands, A and B₁, could be distinguished by cluster analysis, which indicated that they possessed different principal components. While the samples of the same brand but from different batches, B₁ and B₂ were not differentiable, which indicated that the cigarette brand B was in good quality consistence. 2) The load vector angle of 0.86 degrees between puff number of cigarette and CO in mainstream smoke suggested that there existed stronger correlation between the two. The load vectors of puff number with moisture content in mainstream smoke and that of draw resistance with nicotine in mainstream smoke were in an orthogonal state, it indicated their weak correlations. The load vector angle was 148.09 degrees between draw resistance of cigarette and moisture content in mainstream smoke, it indicated a certain negative correlation between them. The correlation analysis between each index could provide reference and theoretical basis for the product design and quality control of cigarette.