Abstract: In this paper, a new Weigh-in-motion(WIM) technology for vehicles using rubber oil pipe is put forward, and a portable WIM system is developed. Because there are several factors that affect the output values of the WIM system, the calculation model using BP Neural Network is set up, and used to carry out data processing. Based on the information fusion technique, recognition model for vehicle classification and expert system were set up. The results of series of tests show that the accuracy of this WIM system meets the requirement of overload detection on most roads. Therefore, this WIM system is an economical and effective tool for vehicle overload detection and traffic data statistics.