The machining quality and detection accuracy of aero-engine blades have a very important influence on their service life of blades. To improve the accuracy of blade detection, a high-precision scanning viewpoint planning method based on structured light is proposed in this paper. Firstly, coarse model data was obtained by coarse scanning under the overall size of the blade, and the field of view was determined according to the camera resolution and acquisition accuracy. Secondly, an improved Angle Criterion algorithm was used to extract the boundary, and the boundary segmentation points were determined according to the boundary coordinates and the range of the visual field. The coarse model was sliced by the section line method for a surface, and the internal segmentation points were determined according to the slice results to complete the uniform segmentation of point clouds. Then, a directed bounding box was established for the segmented point cloud data to obtain the coordinates of the center point, and the normal vector was statistically analyzed to determine the orientation of the main normal to generate the viewpoint coordinates for high-precision scanning. Finally, the surface morphology of the blade was tested and verified. The experimental results show that the average standard deviation of the proposed method is reduced by 0.0054 mm and the collected viewpoint is reduced by 1/3 compared with the viewpoint acquisition result of the supervoxel segmentation, which has good application prospects in the machining inspection of thin-walled blades.