Predicting chemical partitioning to environmental phases continues to be a central problem in environmental chemistry and engineering. The fate and effects of organic chemicals are largely determined by the extent to which they partition from the aqueous phase to organic matter and organism receptors. Hence models for predicting the partitioning of organic chemicals have a long and distinguished history of development and application. For most of this history, the models have been empirical correlations using aqueous, particulate and chemical parameters; the most important of these is the octanol-water partition coefficient. Recently models that are based on more mechanistic formulations have been developed that are replacing regression models using octanol as the central chemical descriptor. The Abraham linear solvation energy relationships (LSER) models explicitly consider hydrogen bonding and van der Waals interactions. LSER models for predicting organic chemical partitioning to natural organic matter and narcotic toxicity to organisms will be presented. Results from an organic chemical partitioning model that uses organic matter molecular structures as a solvent molecule and applies a quantum chemical solvation model will be presented. These newer models are supplanting octanol based prediction models and suggest that a more mechanistic approach to environmental partitioning problems is feasible and likely to be fruitful.