In dependability studies of dynamic systems it is important to assess the probability of occurrence for the events sequences which describe the system evolution or which are critical for the mission of the system or for the humans and environment safety. In this paper we use the probabilistic languages framework in order to realize the quantitative assessment and we start by modeling the system as a finite state automaton. This is ulterior transformed in a probabilistic automaton using the embedded discrete time Markov chain. The determination of the languages afferents at each state of the automaton enable to calculate the probability of occurrence for every events sequence that can be subtract from these languages.