Weigh-in-motion (WIM) systems are designed to capture and record vehicle weights when a vehicle directly passes the WIM measurement region in a normal speed range. WIM systems make the weighing process more efficient than the static weighing system. At present, most types of transportation use containers in international trade. Therefore, in this paper, we propose a novel design of a WIM system for tandem containers, and the radial basis function (RBF) neural network (NN) is used to enhance the accuracy of the weighing result. First, we briefly introduced the hardware structure and the system model of WIM. Next, we used the RBF NN to simulate the WIM model and proposed a self-adaptive algorithm for the center of gravity shifting. The WIM software system was implemented, and the experimental results showed that the error of this system was less than 3%.