In most cases the pharmacological activity of plant extracts is not assigned to single components and often not all active ingredients are known. Approaches other than those considering single compounds only to analyze plant material have proven helpful for a better characterization of extracts in their entirety. In this study extracts of willow bark are analyzed by high-performance thin-layer chromatography (HPTLC) and two different pharmacological tests [the 2,2'-azobis (2-amidinopropane) dihydrochloride reaction and the xanthine/xanthine oxidase reaction] with the help of multivariate data analysis. Described are two models using the results of the chromatographic study of 22 various extracts of willow bark and their pharmacological properties. The chromatographic data are obtained by a special TLC scanner that enables measurement of HPTLC tracks simultaneously in the range of lambda = 200-400 nm. Additionally, the developed models are used to predict the activity of another three extracts of willow bark demonstrating the quality of the model.