With the development of the Internet, online video learning resources are becoming more abundant. However, it is often a problem for users to select learning materials suitable for their own level from many learning videos. Therefore, this study designed an adaptive English learning video recommendation system to recommend the videos suitable for the learner's level based on the learner's English ability through the words collected by the user's learning process via input hypothesis theory and fuzzy theory. According to the experimental results, it is proved that the recommendation system designed in this study can achieve a good adaptive recommendation effect and improve learning effectiveness.