Abstract: In order to reflect driving cycles of the vehicles faster and more accurately and to enhance the adaptability of hybrid electric vehicles to different driving cycles, based on 32 driving cycles of ADVISOR vehicle simulation software, the highest speed, average speed, maximum acceleration, maximum deceleration, average acceleration, average deceleration and parking idle ratio were selected as characteristic parameters. The ranges of characteristic parameter of 4 typical driving cycles were obtained by K-Means clustering analysis. At the same time, the driving cycles database was established under MY SQL(my structured query language) software environment, the programs such as the interface of query and the self-determination were written by JAVA programming language, and the random input driving cycle data was self-determined. The simulated results show that the random input driving cycle data can be accurately inserted into the category to which it belongs. With the continuous expansion of the driving cycles database, the subsequent input query is faster, and the response speed of the system can be improved to a greater extent.