Abstract: Aiming at the problem of surface subsidence caused by large-scale mining of underground resources in mining area， an improved grey model describing the relationship between subsidence and time was proposed by taking Yuan'er Coal Mine of Huaibei Mining Group as the test area and combining differential synthetic aperture radar interferometry （D-InSAR） and grey model （GM （1，1）），which realizes the integration of monitoring and prediction of surface subsidence.The detail steps can be described as follows: firstly， in view of Sentinel-1A satellite and D-InSAR technology， the settlement deformation maps of the study area in different periods from November 16， 2017 to January 27， 2018 were obtained to analyze the process of dynamic subsidence. Then， according to the settlement value of different time series， the semi-parametric GM（1，1） model estimated by compensated least squares method（BGM（1，1）） and compensated least squares method with relative weight（WGM（1，1）） were established to fit and predict the settlement value.The test results show that D-InSAR technology has obvious advantages in dynamic monitoring of land subsidence in mining area， and its monitoring accuracy reaches millimeter level； in addition， BGM （1，1） and WGM （1，1） prediction models make up for the shortcomings of classical GM （1，1） model， the relative error of four test points predicted by WGM （1，1） is 1.99%~26.64%. The study can provide a theoretical basis for dynamic monitoring and subsequent treatment of ground subsidence in mining areas， furthermore， it has certain significance early warning function and reference.

Key words: mining subsidence, D-InSAR, GM（1, 1） model, monitoring and prediction