A Location-Aware Adaptive Vehicle Dynamics System (LAAVDS) is developed to assist the driver in maintaining vehicle handling capabilities through various driving maneuvers. An Intervention Strategy uses a novel measure of handling capability, the Performance Margin, to assess the need to intervene. The driver's commands are modulated to affect desired changes to the Performance Margin in a manner that is minimally intrusive to the driver's control authority. Real-time implementation requires the development of computationally efficient predictive vehicle models which is the focus of this work. This work develops one means to alter the future vehicle states: modulating the driver's throttle commands. First, changes to the longitudinal force are translated to changes in engine torque based on the current operating state (torque and speed) of the engine. Next, the required changes to the throttle to affect these desired changes to engine torque are estimated by developing a linearized, inverse engine model based on the steady-state conditions. These predicted changes to the throttle command are used in the correction stage of the algorithm in which a full non-linear vehicle model yields the resulting longitudinal force response to the predicted change in throttle. Finally, the predictor-corrector control loop is closed by comparing the desired longitudinal force to the resulting force. The performance of the throttle modulation method is exhibited through simulation results on two test cases. Results show that the proposed approach yields a significant overall performance improvement.