In recent years, flipped classroom has become a popular teaching method. Compared with the traditional teaching method, the flipped classroom gives learners and teachers more opportunities to discuss. However, the flipped classroom has also encountered some difficulties. If we do not consider the different learning conditions of each learner when conducting group discussions in the classroom, the learning effectiveness will not be as expected. Therefore, this study uses the Apriori algorithm in association rule analysis to diagnose learning and implement adaptive teaching, hoping to improve the deficiencies in the flipped classroom. This study developed a multimedia learning system applied in the experiment. In pre-class stage, learners were provided with teaching videos, conducted unit tests online, and then used Apriori association rules to analyze the test results for learning diagnose, derive association rules between concepts, and perform adaptive grouping according to learners' test results. Learners will carry out classroom tasks in the class stage, and then implement post-test and post-questionnaire to analyze whether there are significant differences among learners. Finally, we found that using adaptive teaching of flipped classroom combined with concept map learning diagnosis, there were significant differences in research issues such as learning effectiveness, learning motivation, self-efficacy, cognitive load and programming learning attitude. It is hoped that through the results of this study, meaningful contributions can be made in the research field of flipped classroom and adaptive teaching, and it is also hoped that there can be a theoretical basis for scholars who study these fields in the future.