Predictive Quantitative Structure - Activity Relationship (QSAR) models of Anabolic/Androgenic (A/A) activities for the 4,5α-dihydrotestosterone steroid family were obtained by means of multilinear regression using quantum and physicochemical Molecular Descriptors (MDs) as well as a genetic algorithm for the selection of the best subset of MDs. MDs included in our QSAR models allow the structural interpretation of the biological process, evidencing the main role of the shape of molecules, hydrophobicity, and electronic properties. Attempts were made to include lipophilicity (octanol-water partition coefficient) as well as electronic (lowest unoccupied molecular orbital properties and dipole moment) values of the whole molecules in the multivariate relations. It was found from the study that the calculated net charges by semiempirical methods of different atoms in the steroid nucleus [atoms 4 (ring A), 8 (bridgeheads of rings B and C), 11 (ring C) 13 (fusion points of rings C and D), and 16 (ring D)] contribute significantly to binding affinity. The found MDs can also be efficiently used in similarity studies based on cluster analysis. Our model for the A/A ratio (expressed by the weights of the levator ani muscle/ventral prostate in mice) predicts that 2α, 3α-difluoro-methylene-17α-methyl-5α-androstan-17β-ol (13) is the most potent anabolic steroid. By contrast, the 17α-methyl-2β, 17β-dihydroxy-5α-androstane (16) is flagged as the least potent anabolic steroid. The approach described in this report is an alternative for the discovery and optimization of leading anabolic compounds among steroids and analogues. It also gives an important role to electron exchange terms of molecular interactions to this kind of steroid activity.