Abstract: The application of time-series InSAR technology for comprehensive surface deformation monitoring of earth-rock dams can compensate for the limitations of traditional single-point ground monitoring. To address the challenge of analyzing the large amounts of data obtained from time-series InSAR, this study proposes a clustering analysis method for surface deformation patterns and an approach for identifying abnormal deformation areas. First, based on the deformation patterns captured by InSAR monitoring data, hierarchical clustering algorithms are used to partition the dam surface. Then, reverse cloud generators from cloud models are employed to convert the deformation sequences of InSAR coherent points into cloud parameters, which generalize the deformation characteristics of each partition. Finally, the Local Outlier Factor (LOF) is applied to quantify the degree of anomaly for each coherent point within the partitions to identify abnormal deformation areas. Case studies show that the proposed clustering analysis method can efficiently analyze large volumes of InSAR monitoring data, effectively identify abnormal deformation regions in earth-rock dams, and enhance the application of time-series InSAR technology for analyzing dam settlement deformation trends.