In recent years, online learning course has been drawn much attention from educators and researchers through the world. Nowadays, some popular Massive Open Online Courses (MOOCs) in the world provide video courses for learning various subjects and have been successfully applied to many subjects. On the other hand, there are some researchers have indicated that attention state and learning are strongly correlated. However, despite their importance, it is a complicated procedure that when educators want to observe the attention state of each student in online learning course. To help educators effectively observe students' attention state in online learning environments. This study designed and implemented a real-time attention recognition and feedback system, for measuring changes in learner attention state. Finally, to achieve the goals as mentioned above, this study carried out the stability testing, usability testing, and expert evaluation of the real-time attention recognition and feedback system.