Abstract: Traditional mining subsidence monitoring faces many drawbacks, including massive equipment and manpower investment, low density of monitoring sampling points, poor real-time performance, and difficulty in effectively predicting future subsidence evolution. To address these problems, an integrated technical method is proposed for mining subsidence monitoring and prediction, which combines Small Baseline Subset Interferometric Synthetic Aperture Radar (SBAS-InSAR), Whale Optimization Algorithm (WOA) and Long Short-Term Memory (LSTM) network. Taking the Huipodi Coal Mine in Linfen City, Shanxi Province as the research area, 61 scenes of remote sensing images were processed by SBAS-InSAR technology to obtain the annual average subsidence rate and vertical cumulative subsidence from November 2022 to February 2025. The results show that the surface subsidence in the mining area presents obvious spatial heterogeneity. The maximum annual average subsidence rate reaches −81.71 mm/a, and the maximum vertical cumulative subsidence reaches −158.94 mm by January 2025. A WOA-LSTM prediction model is established for eight typical characteristic points with significant subsidence in the study area to predict the subsidence values of the next six periods. The WOA algorithm is adopted to globally optimize the key hyperparameters of the LSTM network, and the sliding window technique is used to construct time series samples. The WOA-LSTM model is compared with the Whale Optimization Algorithm-Gated Recurrent Unit (WOA-GRU) and single LSTM models. The results indicate that compared with the LSTM model, the mean absolute error (MAE) of the WOA-LSTM model is reduced by 64.8%, and decreased by 44.5% compared with the WOA-GRU model. In terms of root mean square error (RMSE), the WOA-LSTM model is 62.5% lower than the LSTM model and 41.8% lower than the WOA-GRU model. The proposed model significantly improves the accuracy of subsidence prediction, enables prospective judgment on the expansion trend of subsidence funnels in goaf areas, and supports mining areas to formulate targeted prevention and control measures in advance.