The Corona Virus Disease 2019（COVID-19）epidemic has led to a shift from offline to online learning in universities, with mobile learning becoming the regular learning norm. However, students exhibit procrastination in the mobile learning process, which greatly affects learning outcomes. In contrast to the traditional classroom, teachers are less able to monitor the online learning process and are unable to do so effectively. Therefore, identifying students’ procrastination behavior in the mobile learning process and improving teaching efficiency have become issues that need attention and solution. Academic procrastination is an avoidant adaptive behavior that not only affects students’ academic performance but also causes stress and anxiety to the procrastinator. An early detection of procrastination and intervention are essential for students to complete their studies. Academic procrastination is mainly identified using subjective scales, which may lead to biased assessment results. In this study, we constructed a mobile academic procrastination recognition model based on a backpropagation neural network, conducted experiments using mobile learning data from 1332 students at a university in China, and evaluated the accuracy of the experiments. The experimental results showed that using students’ mobile learning behavior data to make objective judgments on academic procrastination can avoid the bias of results caused by subjective measurement and improve the objectivity and accuracy of academic procrastination measurement; the recognition accuracy of the mobile learning procrastination recognition model reached 0.992, which significantly improved the accuracy and efficiency of academic procrastination recognition.