Predicting mammalian bioavailability of PAH mixtures from in vitro bioaccessibility results has proven to be an elusive goal. In an attempt to improve in vitro predictions of PAH soil bioavailability we investigated how energetic input influences PAH bioaccessibility by using a high and low energetic shaking method. Co-inertia analysis (COIA), and Structural Equation Modeling (SEM) were also used to examine PAH-PAH interactions during ingestion. PAH bioaccessibility was determined from 14 historically contaminated soils using the fed organic estimation of the human simulation test (FOREhST) with inclusion of a silicone rod as a sorption sink and compared to bioavailability estimates from the juvenile swine model. Shaking method significantly affected PAH bioaccessibility in the FOREhST model, with PAH desorption from the high energy FOREhST almost an order of magnitude greater compared to the low energy FOREhST. PAH-PAH interactions significantly influenced PAH bioavailability and when these interactions were used in a linear model, the model predicted benzo(a)anthracene bioavailability with an slope of 1 and r2 of 0.66 and for benzo(a)pyrene bioavailability has a slope of 1 and r2 of 0.65. Lastly, to confirm the effects as determined by COIA and SEM, we spiked low levels of benzo(a)anthracene into historically contaminated soils, and observed a significant increase in benzo(a)pyrene bioaccessibility. By accounting for PAH interactions, and reducing the energetics of in vitro extractions, we were able to use bioaccessibility to predict bioavailability across 14 historically contaminated soils. Our work suggests that future work on PAH bioavailability and bioaccessibility should focus on the dynamics of how the matrix of PAHs present in the soil interact with mammalian systems. Such interactions should not only include the chemical interactions discussed here but also the interactions of PAH mixtures with mammalian uptake systems.