Salt domes play a crucial role in hydrocarbon storage, underground construction, solution mining, and mineralization. Therefore, deformation monitoring is essential for analyzing the kinematics and impact of salt domes. This study aims to measure the temporal displacements of the Shah-Gheyb salt dome from 2016 to 2019 and from 2020 to 2022 using the New Small Baseline Subset (NSBAS) Interferometric Synthetic Aperture Radar (InSAR) technique and to predict future displacements through machine learning models. A total of 14 data layers, including topography, remote sensing, hydrology, and geology group were used in Machine Learning (ML). Random Forest Regression (RFR) and Support Vector Regression (SVR) models were employed to project displacements in both the East-West (E-W) and Up-Down (U-D) components through 29 scenarios.

In the E-W direction, the salt dome exhibits a displacement rate of 39 mm/year, while in the U-D direction, it varies between −18 and +6 mm/year. ML predictions and SAR interferometry data processing results for the period 2020–2022 were validated using Root Mean Square Error (RMSE) and the correlation coefficient (R). The RFR model demonstrated the lowest RMSE of 1.9 mm for the E-W component, achieving a maximum R-value of 97.3 %. For the U-D component, the RMSE was 2.8 mm, with an R-value of 55.8 %. Evaluation of the predictive performance of the ML models and a comparison of InSAR and ML outcomes indicated that the RFR model predicted displacement along the E-W and U-D directions between 2020 and 2022 with greater accuracy than the SVR. Furthermore, comparing the displacement predicted by the RFR model using SAR interferometry along two perpendicular profiles confirmed the model's precision.