Organic photovoltaic (OPV) compounds hold great promises in providing sustainable, affordable, and renewable energy resources. One of these molecules is a light-harvesting molecular triad, a covalently linked carotenoid-porphyrin-C60 compound that demonstrates an amazing ability to absorb light and produce excited charge-separation states with a giant dipole in its linearly extended configuration. However, it is structurally deformable and forms a broad variety of conformations in solution at room temperature which reduces its quantum efficiency for solar-energy conversion. Such a drawback impedes the real-world application, despite its low manufacturing costs. Another material is a mixture of Boron subphthalocyanine (SubPc) and C₆₀ molecules where, when excited by photons, SubPc can donate an electron to C₆₀. However, as the two molecules form a bulk heterojunction, there are again a broad variety of conformations that have vastly differing charge transfer efficiency. By collaborating with chemists specialized in analytical theories and quantum mechanical calculations, we develop an integrative approach of characterizing the performance of OPV materials with machine learning, statistical physics, and all-atomistic molecular dynamics simulations that address the electronic states of a compound. We aim to instantaneously capture the molecular features for charge transfer while the supramolecules are “in action”. We are currently interested in developing an accurate representation of polarizable force fields for both the ground states and charge-transfer excited states of OPV materials in solution or in condensed states. We showed that in both the triad and SubPc-C60 materials the conformation with the fastest rate has the lowest population which sheds insight into the principle of improving the overall performance for OPV materials. Our work sheds insight into the principle of improving the overall performance for OPV materials.