This paper proposes a novel, intelligent medium access control mechanism for vehicular ad hoc networks. In particular, the minimum and maximum contention window parameters are dynamically tuned based on network observations, concerning all applications' access categories. The simple, generic, and efficient nonlinear control law built, is based on fuzzy logic control and although we focus on vehicle-to-infrastructure (V2I) communication, the proposed scheme can be easily adapted for vehicle-to-vehicle (V2V) communication. We demonstrate, via simulative evaluation of dense and dynamic conditions, that the proposed scheme offers enhanced differentiation among different applications' access categories; therefore, offering sufficient Quality of Service (QoS) in terms of throughput performance, in contrast with the IEEE 802.11p standard MAC protocol we compared against.