In recent years, with the increasing pressure and surface displacement issues in mines, the safety of miners and the production efficiency of mines have faced enormous challenges. This article proposes a new method based on the Convolutional Neural Network-Long Short-Term Memory (CNN-LSTM) model to address this issue, which utilizes LSTM network technology to achieve real-time monitoring and prediction of mining pressure and surface displacement. In the three experiments in the experimental phase, the CNN-LSTM model designed in this article has a Mean Squared Error (MSE) of 0.0441 and a Mean Absolute Error (MAE) of 0.1654 in the prediction of mining pressure. It can be concluded that the data has been significantly improved and the quality of mine safety management has been enhanced. This not only provides strong technical support for mine safety operations, but also brings new possibilities for mine management.