The purpose of the present study was to analyze the online learning activities of students through the data log of LMS through a learning analytical approach and to find out which variables predict academic achievement. After transforming the data left by the learner into a form that can be properly analyzed, the behavior data was set as an independent variable, and the learner’s academic achievement was used as a dependent variable to develop a model for predicting academic achievement. The study subjects were conducted for 55 students who took a course in Educational Technology, a department of education at S Women’s University in Seoul, independent variable data was collected through LMS’s learner log analysis, and the dependent variable, academic achievement, is the final exam score. The research results are as follows. First, a technical statistical analysis was conducted to examine the trend of online activity data of college students. Based on result, the type, frequency, and average of learners’ learning activities and learning behaviors were found. Second, as a result of analysis related to the development of the academic performance prediction model, the number of course accesses, the number of clicks on learning materials, and the reading of posts significantly predicted academic achievement. Third, as a result of verifying the effect of online behavior in the LMS environment on academic achievement through a multiple regression analysis, the model showed 13.0% of the academic achievement. It was within the scope of predictive explanatory power. Based on these results, several suggestions for the follow-up studies were added at the end.