A new series of piperazines, diazepanes, diazocanes, diazabicyclononanes, and diazabicyclodecanes with affinity for the alpha4beta2 subtype of nicotinic acetylcholine receptors were synthesized on the basis of results from a previous computational study. A predictive 3D-QSAR model was developed using the GRID/GOLPE approach (R2 = 0.94, Q2 = 0.83, SDEP = 0.34). The SAR was interpreted in terms of contour maps of the PLS coefficients and in terms of a homology model of the alpha4beta2 subtype of the nicotinic acetylcholine receptors. The results reveal that hydrogen bonding from both hydrogens on the protonated amine and from the pyridine nitrogen to a water molecule as well as van der Waals interactions between the substituent bearing the protonated amine and the receptor is of importance for ligand affinity. The combination of 3D-QSAR and homology modeling proved successful for the interpretation of structure-affinity relationships as well as the validation of the individual modeling approaches.