In this study, we examined efficient shadowing methods for advanced class of Chinese learners of Japanese in terms of pre-task and number of trials. Two factors were set up as independent variables. The first factor was the sensory modality in presentation of the pre-task, consisting of visual or auditory presentation. The second factor was the number of shadowing trials, with three levels: one, two, and three trials. As the pre-tasks, participants were instructed to read Japanese sentences with “whispering-to-oneself" when being presented visually (visual pre-task) or listen to Japanese sentences when being presented auditorily (auditory pre-task). After doing the pre-tasks, the participants shadowed the same Japanese sentences three times, and were asked to orally recall the sentences at the end of shadowing. Fluency and accuracy of Japanese sentences in both shadowing and oral recall were adopted as dependent variables. The results showed that visual pre-task improved oral recall performance after shadowing more than auditory pre-task. It was also suggested that shadowing fluency improved after three trials of shadowing. The visual pre-task may enhance both phonological and semantic memory of the shadowing material, resulting in better playback performance. Based on these findings, the visual pre-task followed by three shadowing sessions may be one of the most effective learning methods of Japanese sentences for advanced class of Chinese learners of Japanese.