With the popularization of MOOC (Massive Open Online Courses), the number of learning resources grows rapidly, which brings challenges in management and annotation. To solve this problem, this study proposes an AlexNet-based design for automated labeling and management of MOOC learning resources. Utilizing the deep learning capability of AlexNet, the system can efficiently and accurately identify and classify various learning resources. After preprocessing, feature extraction and model training, our system achieves superior performance on the test dataset. In addition, we explore the limitations of the method and future research directions. The results of this study show that deep learning techniques are expected to provide an effective solution for the management and annotation of MOOC learning resources. Our AlexNet-based system significantly enhances the efficiency and accuracy of MOOC resource annotation and management.