Predicting whether, how, and to what degree communities recover from disturbance remain majorchallenges in ecology. To predict recovery of coral communities we applied field survey data ofearly recovery dynamics to a multi-species integral projection model that captured key demo-graphic processes driving coral population trajectories, notably density-dependent larval recruit-ment. After testing model predictions against field observations, we updated the model togenerate projections of future coral communities. Our results indicated that communities dis-tributed across an island landscape followed different recovery trajectories but would reassembleto pre-disturbed levels of coral abundance, composition, and size, thus demonstrating persistencein the provision of reef habitat and other ecosystem services. Our study indicates that coral com-munity dynamics are predictable when accounting for the interplay between species life-history,environmental conditions, and density-dependence. We provide a quantitative framework for eval-uating the ecological processes underlying community trajectory and characteristics important toecosystem functioning.