To solve the problems of difficulty in measuring the relative pose of the target spacecraft caused by differ? ent illumination, information distortion and incomplete feature of the image obtained by the visible light camera under complex lighting conditions in space, a binocular vision relative pose measurement method based on image enhancement and arc feature was proposed. An adaptive image preprocessing algorithm based on multi?scale Retinex with chromaticity preservation (MSRCP) was applied firstly to improve the image quality in the environment of spatial dim light and local strong exposure. Secondly, an ellipse detection algorithm based on the edge arc support line segment was applied to ex? tract the arc segment features of the docking ring on the surface of the target spacecraft and fit them into an elliptical con? tour. Finally, a binocular vision camera was used to build a physical simulation platform for relative pose measurement, and a binocular space elliptical conic measurement model was established to calculate the six degrees of freedom relative pose. The relative pose of target spacecraft was achieved in normal and dim light scenes at close range. The experimental results showed that the average error of relative position is better than 20 mm and the average error of relative attitude is better than 0.3° under normal lighting condition, and the average errors of relative position and pose are better than 30 mm and 1° , respectively, under dim lighting condition. The research results provide a reference for the identification and measurement of targets in on?orbit service missions such as close?range space rendezvous and docking, and have techni? cal reference value.