Due to the continuous epidemic in recent years, the traditional teaching mode of history classroom has been gradually replaced by the teaching mode of self-learning classroom. The teaching mode of autonomous learning classroom has become a popular teaching mode in recent years. However, in the autonomous learning classroom under the current history teaching mode, the lecturer cannot always pay attention to the various states of the students. It is also difficult to understand and receive the information the teacher wants to convey in real time. For this reason, human-computer interaction emotion recognition technology has been proposed and developed. In order to construct and realize the teaching mode of self-learning classroom history, this paper studies the emotion recognition technology of human-computer interaction. The research results show that the introduction of human-computer interaction emotion recognition technology into the construction of autonomous learning classroom teaching mode can fully understand students' emotional behavior. It improves the accuracy of students' emotion recognition by 2.67%, enables students to maintain a good learning motivation, and make reasonable plans and arrangements for the historical time and progress of learning. At the same time, it enhances the history teaching intensity and autonomous learning ability, and improves the original single learning mode. By establishing a new teaching-teacher-student relationship, it creates a good and active autonomous classroom atmosphere.