Tailings dams are usually ponds bounded by valleys or surrounding topography to store mining or other chemical industrial waste. On 25 January 2019, the collapse of a tailings dam at the Córrego do Feijao iron ore mine (Brumadinho, Minas Gerais, Brazil) released about 12 million m3 of tailings, killing over 240 people and posing a considerable and ongoing environmental threat. The stability of tailings dam monitoring is very important and in the present paper, a new InSAR (Synthetic Aperture Radar Interferometry) time series approach is proposed to derive ground displacement maps for use in dam safety monitoring. Compared with the other solutions, the unique feature of the proposed method is that: 1) the new Measurement Pixel (MP) selection criteria has the potential to include relatively more accurate MP pixels and build a more robust network, 2) the multi-level grading system makes it possible to add the MP pixels into the main network step-by-step with external control, and 3) the computing efficiency can be improved by strategically reducing the iteration times. The proposed approach was tested on both simulated and real data. Results show that the Simulated Annealing (SA) method normally has a more accurate estimation as compared to the Quasi-Newton (QN) method, despite its longer processing time. Detailed analysis of the displacement maps was conducted to determine the subsidence processes that result from dam construction.