Abstract: To determine the flavor type of cigarette more objectively and scientifically and provide theoretical basis and guidance for cigarette blending, the volatile and semi-volatile substances in cut fillers of cigarettes were extracted by optimized simultaneous distillation extraction (SDE), and the key aroma components were screened out by stepwise regression analysis, so that a model for quantitatively discriminating the flavor type of cigarette was developed. The flavor types of eight cigarette samples were determined by the developed model and parallely compared with their sensory evaluation results based on smell profile method for model validation. The results showed that: 1) Ten key aroma components were screened out, which in the order of their contribution to cigarette flavor type were farnesyl acetone > palmitic acid > 4-cyclopentene-1, 3-dione > benzyl benzoate > myristic acid > dihydroactinidiolide > benzyl alcohol > methyl hexadecanoate > furfural > megastigmatrienone. The regression equation of the said ten components versus cigarette flavor type index Y_quantity was of extremely significant level, it indicated that the developed model possessed high reliability. 2) The flavor types of the eight cigarette samples determined by the two methods were pretty well consistent with each other with the coincidence rate of 86.67%.