Abstract: The purpose of this study is to investigate the role of working memory capacity in L2 Communication in terms of how L2 learners benefit from three different types of corrective feedbacks and the differences among higher and lower working memory capacity levels.Eighty university freshmen in English major were randomly assigned into four groups:recasts group, metalinguistic clues group, metalinguisitic clues+recasts clues group and control group.The study arranged a pre-test, a verbal working memory capacity span test, four one-way information gap treatment tasks, an immediate post-test and a delayed post-test.The results show that WMC moderate the effects of recasts but not metalinguistic clues and metalinguisitic clues+recasts on the acquisition of that-trace filter.Moreover, the higher L2 learners' WMC is, the immediate and delayed mediating effects of WMC on recasts are more obvious.