Abstract: To effectively monitor the quality of blend modules during cigarette manufacturing via near infrared spectroscopy combined with chemometrics,Hotelling T² statistics were derived from the established categorized models for principal components of blend modules and used to monitor the samples before and after blend modifying.The Hotelling T² statistics of 150 samples before blend modification were of 95% confidence interval(Hotelling T²=22.70),the 50 samples of modified blends were of 99% confidence interval(Hotelling T²=28.16).The contents of nicotine,total nitrogen and total sugar in samples before and after blend modification were predicted with the established calibration model,the results indicated that the contents of nicotine,total nitrogen and total sugar in the 150 samples of unmodified blends were within the quality control limits of M±3S,while those in the 50 samples of modified blends all exceeded the limits of M±3S.The method is capable of monitoring the quality consistency and uniformity of blend modules and predicting the variation of chemical component contents in blend modules.