An efficient supercritical fluid chromatography-mass spectrometry (SFC-MS) method was developed for the quality evaluation of Panax Notoginseng (Burk) F.H. Chen (P. notoginseng) by combination with chemical pattern recognition (CPR). Design of experiments (DoE) was applied to obtain optimal SFC-MS conditions. Several CPR methods including hierarchical cluster analysis (HCA), principal component analysis (PCA) and partial least squares discriminant analysis (PLS-DA) were employed to establish a classification model based on the peak areas and contents of 12 components in P. notoginseng in order to evaluate the quality difference according to the collecting time (Chunqi and Dongqi) and medicinal parts (fibrous root, rhizome, branch root, and main root). PLS-DA has proved to be a satisfactory method with accurate discrimination of the selected samples. The characteristic variables based on the variable importance in projection (VIP) values were selected using PLS-DA. Three characteristic components (ginsenoside Rg2, ginsenoside Rg1, ginsenoside Rb1) with higher VIP values (>1) were chosen to further build the CPR model. Subsequently, the model was verified by testing another set of samples and the results indicated that the established model was satisfactory. PLS-DA models based on the peak areas of the 12 selected analytes in 30 batches of P. notoginseng could give accurate classification. The obtained results demonstrate that the developed method using SFC-MS and PLS-DA has a great potential for the quality assessment of P. notoginseng.Copyright © 2022 Elsevier B.V. All rights reserved.