This paper presents a novel approach for classifying microseismic signals using a transfer learning-based convolutional neural network (CNN) called microseismic signals-convolutional neural network (ms-cnn). The ms-cnn model demonstrated superior performance, achieving 99.6% accuracy and processing capabilities that significantly improve recognition timeliness compared to traditional methods. The study emphasizes the potential of deep learning for efficient automatic classification of microseismic events in practical engineering applications.