The study aimed to investigate how working memory (WM) capacity and repetition trials influence shadowing performance of using Japanese for advanced Japanese learners who are Chinese native speaker. Participants were divided into two groups in terms of their WM capacity, both instructed to shadow two Japanese passages for three times. Their shadowing performance was measured by three indices: comprehension and memory performance on Japanese passages, and the accuracy and fluency of oral reproduction. Three main findings were revealed: (1) The result of comprehension performance did not show a significant difference between the participants with high WM capacity and those with low WM capacity; (2) As compared with those with low WM capacity, participants with high WM capacity were better able to memorize the content of the passages, and to orally reproduce the passages with fewer errors and higher fluency; (3) two repetition trials were enough to improve the accuracy and fluency of oral reproduction to a level near the ceiling point. And the difference according to the size of WM capacity was not found.