This paper deals with the diagnostic of functional failure of Cyber- Physical Systems (CPS). Due to the increasing complexity of these systems with heterogeneous technologies, development of more elaborate diagnostic and maintenance techniques are required. An approach is proposed to develop a start-up tests to check the state of health of CPS. We have developed a modeling methodology for CPS based on the Fault detection and isolation models (FDI) combined with finite timed automaton. The proposed method uses supervised learning supplemented by the knowledge of some temporal parameters of the systems. The different steps of our approach are illustrated on the example of a telepresence robot named “RobAIR”.