Abstract: To accurately identify the growing area of flue-cured tobacco, the contents of chemical components, including total sugar, reducing sugar, total nitrogen, nicotine, total chlorine and total potassium, in 402 cured tobacco samples collected from Yunnan, Henan, Anhui, Fujian, Guizhou and Jilin Provinces in 2010 were tested, and the samples were scanned by near infrared spectrometer. The near infrared spectra (NIR) pattern recognition models of growing area were developed by principal component analysis (PCA) and support vector machine (SVM) algorithms, and the growing areas of the samples were recognized. The results indicated that: 1) The prediction accuracy recognized by NIR-PCA-SVM models reached 97%, while that by chemical component-SVM and NIR-SVM models were lower. 2) The NIR-PCA-SVM, and chemical component-SVM models all offered better recoginition for Yunnan tobacco samples. NIR-PCA-SVM model could be applied to pattern recognition of flue-cured tobacco samples of different origins.