Time series analysis has been applied to the field of education, which can evaluate students’ learning behavior and analyze the efficiency of translation, as well as find out the problems in learning. This study takes a university in the central region as the research object and obtains the video learning data in the business English translation teaching course. The four actions of playing, pausing, skipping and completing are analyzed, and Kmeans clustering and statistical methods are used to analyze the learning atmosphere and fluctuations in learning efficiency. The study shows that the clustering analysis of the learning behavior characteristics data yields three types of characteristic behavior results. The learning efficiency of the first class of learning users > the second class of learning users > the third class of learning users. The learning efficiency of the three types of learning users in stage 4 and 5 are in a relatively low state. The learning efficiency of the three types of learning users in the middle stage is higher. Therefore, the method of this paper is feasible in assessing translation efficiency.
