A new approach in reliability evaluation of small isolated hybrid power systems, which include wind turbines and conventional generators, based on fluid stochastic Petri net (PN) modelling is presented. A major novelty of the proposed methodology is that a selection of constant time intervals is adopted, instead of assuming continuous dynamics.. The proposed analysis presents similar characteristics with simulation methods, with the additional advantages of graphical representation of system's components and attributes. Moreover, emphasis is given on the parameterisation of input data that describes operational and reliability characteristics of the system, so a global investigation of the examined system can be done. In order to import more precise data in the simulation process modelled by the PN, the concept of database arcs is introduced. The evaluation is based on the calculation of system's basic indices related to reliability estimation and energy production. The results obtained verify the flexibility and the capabilities of the proposed methodology.