One of the most recent researches in modern diesel engines is fault diagnosis and also alerting drivers in case of occurrence an accident. Thus, an appropriate mathematical engine model should be prepared to establish a base for model based fault diagnosis. Mean Value Engine Models (MVEM) are among modeling strategies that are mainly suitable for their simplicity and mean value estimation of engine state equations. SIMULINK toolbox of MATLAB is utilized as engine modeling software and the results are validated with a real diesel engine. In this work the model based fault diagnosis system is constructed in the framework of Artificial Neural Network (ANN) classifiers which is shown to be an effective tool in diagnosing faults in diesel engines' intake manifold.