According to the characteristics of complex illumination changes in different environments during positioning and tracking,an ORB feature point extraction method with adaptive threshold is proposed. Firstly, the threshold is extracted adaptively by calculating the feature points around the image block automatically,which is then used for extracting the image feature points. Then,the horizontal binocular model is used for feature point tracking.Datasets are used for tests under different environment for comparison of different methods. The experimental results show that the improved method in this paper has larger detection range,higher extraction speed,higher stability in feature point extraction,and has better RMSE performance in trajectory tracking comparison.