Abstract: This study was aimed to establish a fast analysis method for starch and total nitrogen of tobacco according to FTIR spectra. FTIR spectra of tobacco leaf and cigarette were collected and relationship between characteristic vibration skeleton and starch content as well as total nitrogen content were analyzed, and regression mathematical models were established for their fast quantification.Results showed that the area of C-O-H vibration skeleton ( 1339cm- 1) significantly correlated with the content of starch ( p <0.05) and the determination coefficient of regression model was 0.950.1318cm- 1 ( C- N) vibration skeleton area significantly correlated with total nitrogen content ( p < 0.05) and the determination coefficient of regression model was 0.986.Seven unknown samples were used for model's verification.There was no significant difference between the true value and the predicted value calculated by regression models about the content of starch and total nitrogen ( p > 0.05) .Accordingly, FTIR can be used for fast quantification for starch and total nitrogen in tobacco.