This study is aimed to design a novel vocabulary learning mechanism (VLM) in the previously developed video-annotated learning and reviewing system (VALRS) that allows learners to identify unfamiliar or unknown words when listening to the video and generate personalized input enhancement from English subtitles for vocabulary learning. It is expected that the VALRS with VLM (VALRS-VLM) can help learners improve their English listening comprehension in the recognition of speech sounds of individual vocabulary words and meaning understanding of vocabulary words and oral sentences. To investigate the learning effectiveness and learners' experiences of the newly developed VLM, this study compared the outcomes of English listening comprehension performance and technology acceptance between the experimental group using the VALRS-VLM and the control group using the VALRS without the VLM (VALRS-NVLM). Analytical results show that the learners using the VALRS-VLM achieved remarkably better in both English listening learning outcomes and learning retention for vocabulary learning and overall English listening comprehension performance than those who used the VALRS-NVLM. Besides, according to the results of the questionnaire surveyed after the experiment, both the VALRS-VLM and VALRS-NVLM groups showed good technology acceptance. This study confirms that the proposed VALRS-VLM could effectively facilitate learners' English listening learning as well as provide them with a positive learning experience.