In this paper, a dynamic wear model to predict wear volume in a mechanism, involving aleatory and epistemic uncertainty, is established. In this case, harmonic drives are applied to alleviate the impact of clearance for wear of the mechanism. The contact model of a mechanism with clearance that is subjected to harmonic drive is established, with the nonlinear spring-damp model and flexibility of harmonic gear considered. Especially, a slider-crank mechanism with clearance is simulated. The result shows that the cushioning effect of collisions for clearance with application of harmonic drive is superior to that through the flexibility of mechanical parts. Following confidence region method (CRM) for quantification of aleatory and epistemic uncertainty is proposed to analyze the effect of parameter uncertainty for wear volume during the entire time domain, and double-loop Monte Carlo sampling (MCS) approach is improved to propagate uncertainties during the entire time domain. Finally, based on Kriging model, the CRM is used to analyze the effect of parameter uncertainty for wear volume. The result shows that, when both aleatory and epistemic uncertainties are considered, the wear volume boundary is wider and better than that when only aleatory uncertainty is considered. These analyses help to improve the reliability design of system and set a theoretical foundation for the mechanic design and precision analysis of the mechanical system.