The aim of our investigations was to construct activity models of alpha-asarone analogs and related compounds using in treatment of atherosclerosis. The models are based on pseudo- and minireceptor modeling. As a reference set of molecules 20 alpha-asarone analogs were chosen. All of them were flexibly fitted to the four most probable alignments generated from the most active analogs. Subsequent analysis was conducted using pseudoreceptor module Receptor Surface Analysis from Cerius2 package. Statistical analysis based on Genetic Function Approximation revealed several QSAR models. Furthermore, minireceptor program PrGen was applied to the same alignments as above to evaluate atomistic receptor models against virtual particle pseudoreceptor models. Six new compounds with known activities but not included in the QSAR activity model building were used to test the created models.