Amyloid-β (Aβ) protein aggregates through a complex pathway to progress from monomers to soluble oligomers and ultimately insoluble fibrils. Because of the dynamic nature of aggregation, it has proven exceedingly difficult to determine the precise interactions that lead to the formation of transient oligomers. Here, a statistical thermodynamic model has been developed to elucidate these interactions. Aβ1-42 was simulated using fully atomistic replica exchange molecular dynamics. We use an ensemble of approximately 5 × 105 configurations taken from simulation as input in a self-consistent field theory that explicitly accounts for the size, shape, and charge distribution of both the amino acids comprising Aβ and all molecular species present in solution. The solution of the model equations provides a prediction of the probabilities of the configurations of the Aβ dimer and the potential of mean force between two monomers during the dimerization process. This model constitutes a reliable methodology to elucidate the underlying physics of the Aβ dimerization process as a function of pH, temperature, and salt concentration. The results obtained with this new model could be valuable in the design of Aβ oligomerization inhibitors, a prospective therapeutic for Alzheimer's disease.