A similarity-based algorithm based on a previously developed model is applied in the classification of two sets of anticonvulsant and non-anticonvulsant drugs. Each set is composed of a) anticonvulsant compounds that have shown moderate to high activity in the Maximal Electroshock Seizure (MES) test and b) drugs with other biological activities or poor activity in the MES test. The results from the analysis of variance (ANOVA) indicate that the proposed algorithm is able to differentiate anticonvulsant from non-anticonvulsant drugs. The proposed model may then be useful in the identification of new anticonvulsant agents through virtual screening of large virtual libraries of chemical structures.