To address the issue of symmetric ambiguity in multi-view stitching of non-cooperative space targets, this paper proposes a registration strategy based on solar panel alignment. First, the solar panels of the non-cooperative space target are extracted, and an improved Principal Component Analysis (PCA) algorithm is used for initial registration of the panel point clouds. Subsequently, the Iterative Closest Point (ICP) algorithm is employed to further enhance registration accuracy. To mitigate potential symmetric ambiguities in panel point cloud registration, the solar panels are rotated around the three axes of the PCA coordinate system to generate three additional transformation relationships. These transformations are then applied to the current two frames of point clouds. Utilizing the asymmetry of the space target structure on both sides of the panels, the optimal registration result for the current frame is selected from the four registration results. Experimental results demonstrate that the proposed algorithm effectively reduces registration errors caused by symmetric ambiguity and improves the registration success rate of non-cooperative space targets.