A model-based approach to dynamic toll pricing has been developed to provide a systematic method for determining optimal freeway pricing schemes. A novel approach is suggested for alleviating congestion, which utilizes identified models of driver behavior and traffic flow, as well as optimization of the target density to maximize throughput. Real-time traffic information from on-road sensors is integrated with historical information to provide feedback and preview for the dynamic toll price controller. The algorithm developed here provides an opportunity to improve on existing toll policy by guaranteeing minimum speeds for toll lane drivers, maintaining consistent traffic flow for the other drivers, and optimizing the overall traffic throughput.