In the introduction of this paper, we summarize the usual parametric probabilistic models for random uncertainties modeling in dynamic systems and we introduce a new nonparametric probabilistic approach. In a first part, a random matrix theory, recently developed by the author, is introduced allowing the nonparametric probabilistic approach to be constructed. This theory is compared with the usual GOE from the random matrix theory. In a second part of the paper, we present applications of this new theory for linear structural vibrations with dynamical substructuring techniques and for nonlinear transient elastodynamics.