The current competitive and dynamic economic context, requires to manufacturing companies to be reactive and dynamic. For that, they should get a flexible manufacturing system able to change the production plan easily and economically. A good Flexible Manufacturing System (FMS) must have a transportation system able to support a dynamic scheduling and potential failure of the manufacturing system. In this paper, we present a methodology for the dynamic evaluation of the routing flexibility of a manufacturing system. This methodology is based on the computation of entropy and uses Coloured Petri Net model and simulation. We have two objectives:(1) the development of a simulation model able to forecast the level of flexibility of a FMS before its implementation, (2) the definition of an indicator for measuring flexibility. For illustrating our methodology we take an example of a flexible job shop of four machines and one transport resource.