Abstract: This study attempted to use electronic nose system to measure cured tobacco leaf samples of various forms, so that a method for determining volatile component in tobacco leaves could be established. The results showed that the detection results were influenced by tobacco forms and storage periods. Storage time of 4 hours was the best for detecting cut tobacco leaves, and 30 minutes was optimum for tobacco powder. The detection results for cut tobacco was better than those of tobacco powder. Linear discrimination analysis (LDA) discriminate different tobacco materials more effectively than principal component analysis (PCA). The result of LDA could represent overall characteristics of sample. Tested by the Loadings analysis, sensor 2, 7, 8, 9 had higher contribution rates during test cut tobacco sample, while, sensor 2, 7, 9 had higher contribution rates during test tobacco powder. Electronic nose detection could be used to discriminate the varieties or regions produced of flue-cured tobacco leaves.