The aim of reverse pharmacognosy is to find new biological targets for natural compounds by virtual or real screening and identify natural resources that contain the active molecules. To demonstrate the applicability of this concept, we report here a study on epsilon-viniferin, an active ingredient for cosmetic development. Nevertheless, this natural substance is weakly defined in terms of biological properties. SELNERGY, an inverse docking computer software, was used to identify putative binding biological targets for epsilon-viniferin. Among the 400 screened proteins two targets were retained. For cosmetic application, cyclic nucleotide phosphodiesterase 4 (PDE4) was the most interesting candidate. Moreover, other PDE subtypes (1, 2, 3, 5 and 6) were not retained, indicating a selectivity for PDE4. The experimental binding tests on the 6 subtypes of PDE revealed a significant selectivity of epsilon-viniferin for the PDE4 subtype. This selectivity was confirmed by evaluation of epsilon-viniferin on the secretion of TNF-alpha and Interleukin-8. Our data demonstrated that epsilon-viniferin possesses anti-inflammatory properties by inhibiting PDE4 subtype. In conclusion, reverse pharmacognosy and its inverse docking component cannot only be integrated into a program for new lead discovery but is also a useful approach to find new applications for identified compounds.