Land subsidence is a major factor that affects metro line (ML) stability. In this study, an improved multi-temporal interferometric synthetic aperture radar (InSAR) (MTI) method to detect land subsidence near MLs is pre-sented. In particular, our multi-temporal InSAR method provides surface subsidence measurements with high observation density. The MTI method tracks both point-like targets and distributed targets with temporal radar back-scattering steadiness. First, subsidence rates at the point targets with low-amplitude dispersion index (ADI) values are extracted by applying a least-squared estimator on an optimized freely connected network. Second, to reduce error propagation, the pixels with high-ADI values are classified into several groups according to ADI intervals and processed using a Pearson correlation coefficient and hierarchical analysis strategy to obtain the distributed targets. Then, nonlinear subsidence components at all point-like and dis-tributed targets are estimated using phase unwrapping and spatiotemporal filtering on the phase residuals. The proposed MTI method was applied to detect land subsidence near MLs of No. 1 and 3 in the Baoshan district of Shanghai using 18 TerraSAR-X images acquired between April 21, 2008 and October 30, 2010. The results show that the mean subsidence rates of the stations distributed along the two MLs are -12.9 and -14.0 mm/year. Furthermore, three subsidence funnels near the MLs are discovered through thehierarchical analysis. The testing results demonstrate the satisfactory capacity of the proposed MTI method in pro-viding detailed subsidence information near MLs.