Following with previous studies [Tetrahedron (2002), 58, 4383; Tetrahedron (2004) 60, 9417], the present work intended to interpret the dissociation process of model insoluble peptide sequences in the light of their secondary structures and the electron acceptor (AN) and electron donor (DN) solvent system properties. The (1-42) DAEFRHDSGYDVHHQKLVFFAEDVGSQKGAIIGLMVGGVVIA b-amyloid peptide involved in the Alzheimer's disease, its (1-21) DAEFRHDSGYDVHHQKLVFFA fragment and the well known strongly aggregated VVLGAAIV were selected as models of insoluble peptides and studied through solubility and CD experiments. Solvents presenting similar AN and DN values such as acetonitrile and acetone failed, despite their polarities, to dissociate peptide chains (free in solution or even bound to a polymer). This is due to a self-neutralization effect occurring in these solvent molecules characterized by presenting rather similar AN and DN numbers. Otherwise, the maximum solubility of these aggregated sequences was attained in solvents presenting the highest possible (AN-DN) values (in positive or negative mode). The AN-DN values ranged from approximately −20 (DMSO) to +80 (HFIP) and, notably, the lowest dissociation power was ascribed to solvents presenting values of approximately +40. The strong electrophilic water is located in this region, indicating that, for dissociation of specific insoluble segments, the solvent should appropriately combine its acid/base strength with the potential for van der Waals interactions. In this context, the water addition for solubilization of associated peptides is only beneficial when in TFE but not in strongly nucleophilic solvent DMSO. We also observed a sequence-dependent pH effect on peptide solubility confirmed through CD spectroscopy. This experimental approach also revealed a complex but, in many cases, consistent influence of peptide conformation on its solubility degree, observed with the use of structure-inducing solvents such as TFE or acetronitrile. To date, no clear rules have been proffered for the way in which solvent systems are selected for a given experiment, and this process continues to be conducted in a random fashion. From our perspective, the results of the present study reveal some consistent guidelines for addressing this problem, primary based on the Lewis acid-base concepts and with a clear potential for application in the solubilization prediction of many other physiologically relevant macromolecules. Supported by FAPESP and CNPq.