To overcome the limit of morphological method for classification of herbal drug, a novel method to discriminate its origin using pyrolysis mass spectrometry-multivariate analysis was developed. This method was applied successfully to Scutellaria baicalensis Georgi, one of the most popular herbal drug in oriental countries. The ethylacetate soluble fractions were prepared by sonication from pulverized roots of S. baicalensis which were collected from various regions including Korea and China, and subjected to direct insertion probe (DIP) mass spectrometry to achieve mass spectra of pyrolizates of extracts. The probe temperature was elevated from $30^{circ}C$ to $320^{circ}C$ at increasing rate $64^{circ}C/min$, and the average mass spectrum calculated from total ion chromatography (TIC) was obtained. The relative peak intensities versus m/z were subjected to SAS program, and the training set (9 from Korea origin and 22 from China origin) was clustered two groups as its origin. In the test set, 11 samples among total 13 test sample were successfully classified according to their origin by developed method with accuracy of 85%.