Abstract: In order to solve the problem of location error, which caused by multipath interference and time delay in NLOS environment of mine based on location method of signal receiving intensity (RSSI) and time of flight (ToF), a binocular vision tracking and positioning method for mine moving target based on ORB feature is proposed.Firstly, through establishing the binocular vision tracking and positioning model of mine moving target, using the RSSI threshold to monitor the position range of underground moving target, the binocular vision sensor is triggered to do the image information sampling and feature extraction of moving target.After that, the target image and the training image are matched by the ORB algorithm.Finally, according to the parallax mean and the pixel coordinate formula of binocular vision sensor, the coordinates of the camera's internal and external parameters and the coordinates of the moving target are calculated.Therefore, the accurate location information of the mine moving target is obtained to realize a real-time tracking and accurate positioning.The results show that, compared with other algorithms, the proposed method can effectively improve the positioning accuracy and real-time performance, and has strong robustness to the tracking and positioning of moving objects under the target occlusion, low illumination and noise environment.