Abstract: By integrating wear prediction with flexible multibody dynamics, the wear at the revolute clearance joint in flexible mechanical systems was predicted. The flexible component was modeled based on the absolute nodal coordinate formulation (ANCF), and the contact force in the clearance joint was calculated using continuous contact force model. An iterative wear prediction procedure based on Archad's wear model was employed. In order to determine wear coefficient, pin-on-disc experiments were performed, and Radial Basis Function Neural Network (RBFNN) technique was introduced to deal with tests data to obtain the wear coefficient at different contact conditions. Finally, a planar slider-crank mechanism with clearance joint was used to perform numerical simulations. The results show that after taking components' flexibility into account, the impact value in the clearance joint was much lower than that in the case without considering such factor, and the wear result predicted was also slightly reduced. Moreover, this difference was more pronounced when the simulation time was longer.