A new approach to apply dynamic relationships in dynamic Bayesian networks is presented. This model makes use of some of the concepts of functional event sequence diagrams for modeling dynamic relationships between component failures, physical variables, and measurements, and presents the required equations for derivation of conditional probability values to be automatically mapped into a temporal Bayesian networks. Using this approach, we can exploit expert knowledge for developing temporal BN and elicitation of required probability values more efficiently. The developed model using block diagrams will be also more informative and expressive. An application of this methodology is also presented for taking dynamic condition into account for diagnosis and prediction analysis.