An enduring challenge in the understanding of internal protein motions is the effective separation and characterization of diffusive and vibrational dynamical components. To address this problem, here nanosecond molecular dynamics trajectories of myoglobin in aqueous solution, performed over a range of temperatures between 120 and 300 K, are subjected to principal component analysis, and the coordinate autocorrelation functions of the resulting principal modes are interpreted using a model combining damped Langevin vibration within potential wells and barrier-crossing diffusion between them. Both the vibrational frequency and the fraction of the mean-square fluctuation arising from vibrational motion undergo transitions with temperature at about 180 K. In contrast, the vibrational friction remains linear with temperature. The diffusional component of the mean-square fluctuation increases dramatically at the dynamical transition. The heights of the energy barriers between the potential wells are estimated, and the associated diffusion constants are calculated using Kramers' rate theory. Model functions of the frequency dependence of the frictional and diffusional quantities are obtained. The dynamic structure factor from the full molecular dynamics trajectory is well reproduced by the model. Overall, the results indicate that a global description of nanosecond temperature-dependent diffusion and vibrational internal protein dynamics can be obtained by applying the results of the present diffusion-vibration model to the vibrational motions obtained from a normal-mode analysis.