Intrinsically disordered proteins (IDPs) access highly diverse ensembles of conformations in their functional states. Although this conformational plasticity is essential to their function, little is known about the dynamics underlying interconversion between accessible states. Nuclear magnetic resonance (NMR) relaxation rates contain a wealth of information about the time scales and amplitudes of motion in IDPs, but the highly dynamic nature of IDPs complicates their interpretation. We present a novel framework in which a series of molecular dynamics (MD) simulations are used in combination with experimental (15)N relaxation measurements to characterize the ensemble of dynamic processes contributing to the observed rates. By accounting for the distinct dynamic averaging present in the different conformational states sampled by the equilibrium ensemble, we are able to accurately describe both dynamic time scales and local and global conformational sampling. The method is robust, systematically improving agreement with independent experimental relaxation data, irrespective of the actively targeted rates, and suggesting interdependence of motions occurring on time scales varying over 3 orders of magnitude.