Large-scale mining of coal resources has caused surface subsidence, threatening the ecological environment and surface safety and hindering the sustainable development of coal resource-based cities seriously. In order to identify coal mining subsidence area more accurately and quickly, a novel method for automatic identification of coal mining subsidence area was proposed in this study, which is based on InSAR technology and time series classification algorithm. The method comprehensively considered the quality of data acquisition and the degree of automation of algorithm implementation, realizing automatic identification of large-space coal mining subsidence area in the case of small samples. Firstly, the unclassified subsidence dataset is constructed by SAR data around time series combined with InSAR technology. Secondly, the degree of change of time series subsidence is used as the basis of classification, and it is divided into two categories: regular type of coal mining subsidence and irregular type of non-coal mining subsidence. Finally, Dynamic Time Warping (DTW) algorithm is used to calculate the similarity distance between the time series curves of each image element of the subsidence dataset to be tested and the standard coal mining subsidence type curves, which aim at identifying the coal mining subsidence area. Through utilizing the multitude filtering algorithm to filter the recognition results, the automatic recognition results of coal mining subsidence area is finally obtained. The results show that this method can still be used to efficiently identify and circle coal mining subsidence area with only 640 samples, and the average Overall Accuracy (OA) and Kappa Coefficient (KC) of coal mining subsidence area are 0.91 and 0.83, respectively. The method can not only identify the existing and potential subsidence area in the mining area, but provide technical support for ecological environment restoration and early warning of subsidence damage.