The dynamic behaviour analysis is to investigate the interferences and dependences among components or subsystems within a redundant system. Due to insufficient historical date, complex system structures, human factors, complex working environment etc., parametric uncertainty should also be taken into account when people model dependences. In this paper, we present a dynamic reliability model for common causes failures of k-out-of- N:G redundant systems, which is not only to demonstrate failure process of redundant systems, but also to concern parametric uncertainties in lifetime distributions of components or subsystems within the systems. In the presented model, the parametric uncertainties of lifetime distributions are expressed in the form of probability box firstly; homogeneous continuous-time Markov chain is then applied to represent the failure process of a k-out-of-N:G redundant system, in which occurrence of a common cause failure between two components is concerned. Finally, this model is applied to an auxiliary power group of an eight-rotor unmanned aerial vehicle (UAV) system to verify its effectiveness. Keywords: Common cause failure, Parametric uncertainty, Probability box, Homogeneous continuous-time Markov chain, k-out-of-N:G redundant system.