The service level of the tollplaza is an important guideline to describe the characteristic of toll road, and it’s also a Scientific basis to design the tollplaza and to manage the toll collection system. In view of the difficulty and complexity of a mathematical and physical method to classify the service level and the good behavior in self-organization and robustness of Kohonen neural network, this study applies Kohonen neural network to classify the service level of ETC tollplaza. In this paper, the method is applied to classifi the service level for a certain tollplaza with five toll lanes, and it’s found to bepotentially applicable in practice.