In this paper, we propose an approach allowing us to study the performance of a mechatronic system using Petri Nets in which we use particularly the PH12 method for the evaluation of failure probability of a mechanical component. The Petri Nets are used to describe the process of sequential dynamics of a system. Deterministic stochastic Petri networks have been implemented to introduce random phenomena following the laws of distribution. After the functional and dysfunctional modeling of a mechatronic system using Petri Nets, it is necessary to associate times to different transitions of the network. It is necessary to take into account a physical modeling in order to determine functional times of different components which are associated with functional transitions. Regarding dysfunctional transitions, the laws of probability are assigned depending on the knowledge concerning the reliability of different types of components: the exponential law for electronic devices, the Jelinski-Moranda model for software, etc. For the mechanical components, we use the PH12 method for the evaluation of the time variant failure probability.