This paper presents development and application of a real-time virtual differential map-matching algorithm. Taking advantage of the slowly drifting property of the GPS errors and the continuity in map errors, the authors first define a differential vector between the vehicle's GPS track points and their corresponding mapped points by historical tracks. This information is then applied to a pre-correction process for examining the on-line vehicle tracks. As the differential vector is only the approximate of the real-time deviation (i.e., virtual differential vector), it is corrected continuously along with the vehicle movement. As a result, the algorithm takes into account both historical information and real-time data of vehicle velocity and direction, as well as the topological structure of the road network map in the map-matching process. The performance of the algorithm is examined through a sample application.