A method of comprehensive chemical pattern recognition of Atractylodis Rhizoma was established by GC-MS fingerprint, principal component analysis, cluster analysis and discriminant analysis. A DB-wax column (0.25 mm x 60 m, 0.25 microm) with El ion source and 70 V electron multiplier were used for GC-MS analysis. Using principal component analysis, cluster analysis, and discriminant analysis, 15 common peaks of sample fingerprints for chemical pattern recognition research were analysed. The same results were obtained from the fingerprint, principal component analysis and cluster analysis, which could use to distinguish genuine Atractylodes lancea, ungenuine A. lancea and A. chinensis. Thus, this method could be used for the quality control and comprehensive evaluation of Atractylodis Rhizoma.