This paper investigates the use of the probabilistic and continuous extensions of action systems in the development and calculation of reliability of continuous, real-time systems. Rather than develop a new semantics to formally combine the existing extensions, it investigates a methodology for using them together, and the conditions under which this methodology is sound. A key feature of the methodology is that it simplifies the development process by separating the probabilistic calculations of system reliability from the details of the system’s real-time, continuous behaviour.