Abstract: Near infrared spectroscopy (NIRS) in combination with the chemometric methods of principal component analysis (PCA) and Mahalanobis distance was applied to the study of quality uniformity of cut tobacco samples collected at the 4 process sections (i.e., blending, flavoring, storaging and feeding) in the course of technological processing of cut tobacco for making cigaretts. The characteristic informations contained in the NIR spectra extracted by PCA were used to reveal comprehensively deviations between samples taken from the 4 sampling points; and fluctuations of quality of cut tobacco during its processing were revealed by the data of Mahalanobis distance. In addition, by using the model extablished by finished products of tobacco, contents of 6 chemical components (i.e., total plant alkaloids, total saccharides, reducing sugar, total nitrogen, potassium and chlorine) in cut tobacco samples collected at sampling point D can be predicted, and difference of contents of the chemical components in cut tobacco samples of different batches can be controlled in a reasonable ranges, in order to lower the fluctuation of contents of the chemical components in products of cigarettes, and to keep a stable characteristic style for a name-brand cigarette.