The surface subsidence, with a maximum value exceeding 160 m, has developed over 40 years due to underground magnetite mining. The classical mapping and sensing method was proven to be difficult to monitor this progressive surface subsidence. This study aims to address the issue of surface subsidence associated with underground mining by using UAV photogrammetry technology to track its evolution. The UAV has flown 11 times at equal intervals over 1 year. The subsidence data were analyzed using a proposed time-dependent model, which describes the deflection with respect to time as linear segments of different dimensions in log–log scales. Besides, the exponent of the model is represented by the slope ratio of these linear segments. The proposed model was calibrated using existing surface subsidence data to predict future trends. The surface subsidence was to slow down and came to be stable. Results demonstrated that the UAV photogrammetry technology could provide a comprehensive understanding of land space ecology.