The perception of thousands of odours by about 380 human olfactory receptors (ORs) results from a combinatorial coding, in which one OR recognizes multiple odorants and odorants are recognized by different combinations of ORs. We used data of 95 odorant molecules tested on two human ORs, OR1G1 (class II) and OR52D1 (class I), to perform a 3D molecular modelling study of ligands using Catalyst/HypoGen software. The sorting-out procedure previously used for OR1G1 ligands was transposed to OR52D1 ligands. The 3D-QSAR models obtained for OR1G1 and OR52D1 ligands essentially differ by distance between their features. This result suggests a path to decipher the odotopes of OR1G1 and OR52D1 agonists in order to investigate the role of ORs in olfactory coding.