Traditional Chinese medicines are regarded as promising source for drug development, since they are relatively safe and their therapeutic effect are previously known, thanks to accumulated trial experiences for over-thousand years. However, its integrated mechanism of action in multi-component level is still unknown.In this research, we developed a system that predict the effect of the multicomponent drug based on multi-target analysis of molecular information. We have integrated biological databases (KEGG, ChEMBL) to obtain drug-related pathways. Next, pathway entities were divided into more specifically defined smaller sub-pathway named product oriented sub-pathway (PSP). From PSP-linked to metabolite sets, drug-affected terms in physiological level were obtained through co-occurrence analysis in PubMed.We have performed a case study for the complex herbal drugs composed of Corydalis tuber and Pharbitis seed. Resulted physiological function terms showed significant similarity with previously known functions of the TCM ingredients.