Automatic Vehicle Classification becomes more so essential to the electric toll collection of highways and other related applications in intelligent transportation system. A novel and robust vision-based vehicle classification algorithm for highway toll is proposed in this paper. The algorithm focuses on a vehicle?s length, height and width dimension parameters detected simply by a single camera. Multi-sensors are utilized for aided estimating the vehicle, which are well complementary equipments for classifying the passing vehicles. And experimental results show that the proposed algorithm is valuable and correct.