This paper utilizes the GA-FNN algorithms to extract the unusual problems in real-time vehicle monitoring. GA algorithm has the features of searching the global information, thus in this way a new approach is given to select the effective sensor information, and use it as the input of fuzzy neural network. The fuzzy matrix trained by fuzzy neural network can detect the degree of abnormal through real-time operation of the vehicle. The simulation examples demonstrate the validity of the application of the GA-FNN, the abnormal degree can be detected quickly.