In order to monitor the driving status of vehicles in real time, a classification model of vehicle driving safety features based on Canopy-K-means algorithm was established. Using Canopy-K-means clustering algorithm to mine and analyze vehicle driving data, using Euclidean distance as the similarity classification index between attributes of data set, to obtain offline clustering centroids representing different driving safety features; building TruckSim and Simulink joint simulation Platform, set the fixed radius variable speed and steering wheel oblique step input simulation working conditions to identify the vehicle's driving status online; at the same time, in order to verify the application effect of the method on the real vehicle, the same working conditions were set to verify the offline clustering centroid analysis. The simulation and real vehicle results show that the data mining method based on the Canopy-K-means algorithm can classify different driving state data, and the obtained cluster centroids representing different driving safety characteristics can evaluate the driving stability of the vehicle to a certain extent, which can be used as the judgment basis for vehicle control and early warning.