The effect of levodopa in alleviating the symptoms of Parkinson's disease is altered in a highly nonlinear manner as the disease progresses. This can be attributed to different compensation mechanisms taking place in the basal ganglia where the dopaminergic neurons are progressively lost. This alteration in the effect of levodopa complicates the optimization of a drug regimen. The present work aims at investigating the nonlinear dynamics of Parkinson's disease and its therapy through mechanistic mathematical modeling. Using a holistic approach, a pharmacokinetic model of levodopa was combined to a dopamine dynamics and a neurocomputational model of basal ganglia. The influence of neuronal death on these different mechanisms was also integrated. Using this model, we were able to investigate the nonlinear relationships between the levodopa plasma concentration, the dopamine brain concentration, and a response to a motor task. Variations in dopamine concentrations in the brain for different levodopa doses were also studied. Finally, we investigated the narrowing of a levodopa therapeutic index with the progression of the disease as a result of these nonlinearities. In conclusion, various consequences of nonlinear dynamics in Parkinson's disease treatment were studied by developing an integrative model. This model paves the way toward individualization of a dosing regimen. Using sensor based information, the parameters of the model could be fitted to individual data to propose optimal individual regimens.