This paper uses multiple regression analysis to construct a prediction model of online learners’ academic performance. The English online classroom learning data of 448 college students in college B are taken as the object of analysis. From them, four indicators, namely, the number of study times, daily study time, daily study frequency, and task point completion, were selected as independent variables, and learning achievement was taken as the dependent variable. The regression coefficients of each variable were determined by stepwise multiple regression, and the final regression model of academic performance was determined as \(Y = -8.632 + 0.99x_1 + 0.231x_2 + 0.485x_3 + 0.286x_4\) . The results of the constructed prediction model of online learners’ academic performance are basically accurate through the discrimination of sample independence, residual normality test, and the discrimination of the absence of multiple covariance in the independent variables, respectively. Teachers can formulate teaching plans and carry out personalized tutoring according to the English online classroom learning data, and verify the learning effect by comparing the before and after data.
