Abstract: The process quality of cigarette blend was characterized by principal component analysis-Mahalanobis distance in multivariate analysis and partial least square algorithm combined with near infrared spectrometry.The results showed that the method was able to make full use of the information to verify the stability and uniformity of blending process, and promptly and simultaneously monitor the quality information of multi-components in the process.It was practical for the control and analysis of process quality and the monitor of product quality.