<p>This study presents a new approach for monitoring the deformations of reservoir dams by combining the advantages of BeiDuo satellite data and Sentinel-1 SAR data. The model's core relies on Deep Neural networks with Long-Term and Short-Term Memory (DNN-LSTM) models. Traditional monitoring techniques like InSAR and GNSS face limitations regarding data gaps, atmospheric interference and topographical noise. To overcome these issues, the proposed model combines the advantages of high-precision real-time positional data from BeiDou with deformation data from Sentinental-1 SAR, which helps to provide more accurate monitoring solutions. Similarly, the dam's non-linear, time-dependent deformation patterns are captured by the DNN-LSTM model, which is particularly impacted by sudden climate changes like water level variations and heavy rainfall. This fusion ability of the model will effectively predict the long-term deformation trends of the dams, even in regions with low ground-based sensor networks. The simulation of the model is performed under Xiaolangdi Dam using 62 scenes. The study's experiments show that the proposed model effectively outperforms traditional monitoring techniques in monitoring accuracy. Based on these outcomes, the model effectively contributes to future dam deformation predictions by providing adequate decision-making support for engineers and other dam-based constructors to protect the dam infrastructure from sudden climate fluctuation risk.</p>