Organophosphate (OP) nerve agents such as sarin, soman, tabun, and O-ethyl S-[2-(diisopropylamino) ethyl] methylphosphonothioate (VX) do not react solely with acetylcholinesterase (AChE). Evidence suggests that cholinergic-independent pathways over a wide range are also targeted, including serine proteases. These proteases comprise nearly one-third of all known proteases and play major roles in synaptic plasticity, learning, memory, neuroprotection, wound healing, cell signaling, inflammation, blood coagulation, and protein processing. Inhibition of these proteases by OP was found to exert a wide range of noncholinergic effects depending on the type of OP, the dose, and the duration of exposure. Consequently, in order to understand these differences, in silico biologically based dose-response and quantitative structure–activity relationship (QSAR) methodologies need to be integrated. Here, QSAR were used to predict OP bimolecular rate constants for trypsin and α-chymotrypsin. A heuristic regression of over 500 topological/constitutional, geometric, thermodynamic, electrostatic, and quantum mechanical descriptors, using the software Ampac 8.0 and Codessa 2.51 (SemiChem, Inc., Shawnee, KS), was developed to obtain statistically verified equations for the models. General models, using all data subsets, resulted in R ² values of .94 and .92 and leave-one-out Q ² values of 0.9 and 0.87 for trypsin and α-chymotrypsin. To validate the general model, training sets were split into independent subsets for test set evaluation. A y-randomization procedure, used to estimate chance correlation, was performed 10,000 times, resulting in mean R ² values of .24 and .3 for trypsin and α-chymotrypsin. The results show that these models are highly predictive and capable of delineating the complex mechanism of action between OP and serine proteases, and ultimately, by applying this approach to other OP enzyme reactions such as AChE, facilitate the development of biologically based dose-response models.