Studies of 3D models for cyclopentapeptides (CPP's) employing only NMR spectroscopy encounter a serious problem. Because of conformer averaging, 3D structure(s) derived directly from NMR data may not correspond to the energy minimum (minima) with low relative conformational energy. At the same time, independent energy calculations can determine all low-energy conformers For the CPP backbone. The two approaches are compared in this study by results obtained for cyclo(D-Pro(1)-Ala(2)-Ala(3)-Ala(4)-Ala(5)). Contrary to the conclusion (predominance of the beta II'gamma type conformer) of earlier NMR studies, independent energy calculations found a different family of low-energy 3D structures that are consistent both with the NMR data in DMSO and with the known X-ray data on CPP's. The preferable Ala(4) conformations were found in the alpha(R)/alpha(L) regions suggesting studies of cyclo(D-Pro(1)-Ala(2)-Ala(3)-Aib(4)-Ala(5)) which was synthesized. Further NMR studies confirmed the results of the independent energy calculations, The independent energy calculations have been applied also to cyclo(Arg(1)-Gly(2)-Asp(3)-D-Phe(4)-Val(5)) and cyclo(Arg(1)-Gly(2)-Asp(3)-Phe(4)-D-Val(5)). Both peptides are almost equally potent inhibitors of binding of alpha(IIb)beta(3) integrins to fibrinogen and of alpha v beta(3) integrins to vitronectin. If both of them possess a NMR-predicted conformer of the beta II'gamma type, however, the conformations of the active sequence, Arg(1)-Gly(2)-Asp(3), should be dissimilar in these two peptides, This discrepancy is eliminated in the 3D pharmacophore model proposed by independent energy calculations. The model is also in good agreement with the model by other authors that was confirmed by X-ray studies.