Chemical pattern recognition was applied to the quality assessment of the traditional Chinese medicine "wei ling xian". The dried roots and rhizomes of Clematis chinensis Osbeck, Clematis hexapetalo Pall, and Clematis manshurica Rupr, 21 samples of "wei ling xian" of six different species which were collected from different regions of China, were extracted with methanol. The extracts were analyzed by gas chromatography and the GC data of "wei Ling xian" samples were correlated with their anti-inflammatory activity in pharmacologic experiments. Positive anti-inflammatory activity is regarded as the quality assurance of medicinal "wei ling xian". According to the results, the samples were classified as two groups, one having and the other not having distinct anti-inflammatory activity at the 5% significance level. Some samples were used as the training set, while some as the test set. The SIMCA program was used for the extraction of features and classification, whereas the PCA program for displaying the data in a two-dimensional space. The quality of unknown samples of "Wei Ling Xian" can be assessed as a result of the classification or the two-dimensional display.