In recent years, English instruction has emphasized the importance of listening comprehension; however, playing CDs in class is still the most widely used approach to practicing English listening skills in Taiwan. It is obvious that this training approach cannot satisfy individual needs, thus reducing the effect of promoting English listening comprehension performance. Thus, the study proposes a novel video-annotated listening review mechanism (VALRM) with two review modes -- the delayed and immediate review modes to enhance learners' listening comprehension performance. This study examines the effects of learners in the experimental group using the VALRM and those in the control group using Youtube with self-determined listening review mechanism (SDLRM) on English listening comprehension performance. Analytical results show that the learners with the VALRM support achieved significantly better listening comprehension performance than those of using the SDLRM. More importantly, the experimental results confirm that the learners in the experimental group had significantly better listening comprehension abilities even though removing the VALRM support than those of the control group removing the SDLRM support. Additionally, the interview results show most of the interviewees agreed that using the VALRM is more effective in promoting English listening learning performance than using the SDLRM because they could find out the unclear parts and replay them easily and efficiently. The main contribution of this study is to present an effective computer-assisted English listening learning tool based on the two review modes provided by the proposed VALRM in an autonomous learning environment.