Automated technological devices need diagnostics. At present, it is possible and necessary to use a wide range of sensors, computing systems and artificial intelligence methods based on neural networks, fuzzy inference systems in diagnostic systems. This paper proposes three models of information processes for diagnosing mechatronic systems - parallel, sequential and combined. To diagnose the nodes of mechatronic systems showed that it is necessary to combine neural networks and fuzzy logic. For a fuzzy inference system, it is proposed to use such parameters as the value of the diagnostic parameter, the trend of this parameter, and the value of the node operating time. Examples of diagnostics of such units of mechatronic systems as bearings and electric drives based on a neural network and fuzzy logic are given. Experiments have been carried out confirming the correctness of the reliability of the diagnostic model.