本文论述了应用时序分析方法建立车用发动机正常状态常规参数时序模型，并用信息论的Kulback－Leibler（K－L）信息距离判别函数对发动机的运行状态进行监测。试验结果证明了其有效性，可提前数小时对发动机的异常状态发出预报，对车用发动机的故障诊断具有极高的参考价值。 This paper discusses the application of sequential analysis method to establish the normal parameters of vehicle engine normal time series model, and uses the Kulback-Leibler (K-L) information distance discriminant function of information theory to monitor the engine operating status. The test result proves its effectiveness. It can forecast the abnormal state of the engine several hours in advance and has extremely high reference value to the fault diagnosis of the vehicle engine.