To predict the failure modes of a Diesel engine we have to use several condition monitoring techniques. Besides the trend analysis of engine performance parameters, the technique of lub oil condition monitoring is widely used for warning against failure in Diesel engines. It seeks to improve engine predictive maintenance by increasing availability, reliability, reducing maintenance cost and failures in operation. The technique presented in this paper entails a regular and periodic sampling of the engine lubricant oil followed by a laboratory analysis of its condition degradation and contamination. The analyzed data is then stored in a computerized data base for rapid manipulation and graphical retrieval to identify abnormal trends that characterizes the development of a future failure of engine component. Engine diagnosis is then presented by an Expert System.