Abstract: Objectives: The conventional dam deformation monitoring method has been demonstrated to be capable of reflecting the displacement of typical dam sections and elevations. However, the density of measuring points is often inadequate, with important parts consequently being missed. Ground-based synthetic aperture radar interferometry (GB-InSAR) has emerged as an effective means of monitoring regional deformation of dam surfaces. Nevertheless, the precision of long-term sequential monitoring is frequently compromised by atmospheric disturbances, which significantly impedes the ability to obtain reliable deformation measurements through single-master time series InSAR analysis. Methods: In order to address this issue, improvements were made to the GB InSAR time series technology. The optimized approach used continuous imagery acquired during periods of stable meteorological conditions to compute average image sequences, which were then used to construct short temporal baseline interferometric pairs. Atmospheric phase correction was then applied to the interferograms using high-quality pixels (HQP) identified in stable regions, enabling precise deformation inversion. The method was applied to the long-term deformation monitoring of the Baihetan ultra-high hyperbolic arch dam. A comparative analysis was conducted between the timeseries GB-InSAR results and the traditional monitoring method, based on the accurate identification of HQPs across the dam surface. Results: Except for localized regions affected by the sidelobe effect of deep outlets, the deformation sequences derived from both methods at multiple plumbline observation sites exhibit a high degree of consistency, with root mean square error (RMSE) values of approximately 2 mm. Furthermore, the deformation center of the dam, as identified by the GB-InSAR technique, is located at the top of the 16th dam section, which is consistent with the results obtained from numerical simulations. Conclusions: The proposed method has the capacity to overcome the interference of severe atmospheric variations in near-water regions and address the issue of diurnal deformation signals being mixed with atmospheric phases in GB-SAR, which facilitates high-precise measurements of long-term dam surface deformation. The approach provides critical technical support for safety analysis and structural deformation interpretation of high arch dams.