Thermal barrier coating (TBC) is a critical high-temperature protection technology applied to hot-section components of military and civilian aero-engines. Composed of a ceramic oxide top layer and a metallic bond coat, it significantly reduces the substrate operating temperature and offers technical advantages such as high hardness, excellent stability, resistance to high-temperature corrosion, reduces fuel consumption, and improves engine efficiency and service life. After initial preparation via atmospheric plasma spraying, the surface roughness of the coating is relatively high (approximately R_a10 μm). In production, grinding and polishing post-processing are commonly employed to reduce it to the required range (around R_a1.6 μm). This study proposes a spline path curve generation and feature point sampling method based on the RANSAC segmentation principle, applicable to robotic automated grinding and polishing of TBC on small turbine blades in aero-engines. The method utilizes a 3D vision sensor to scan the blade surface in real time, generating point cloud data. Through point cloud processing and B-spline curve fitting algorithms, a full-coverage grinding and polishing path for the high-temperature coating on the aero-engine blade surface is generated. Experimental verification demonstrates that this method reduces the coating surface roughness to below R_a0.7 μm while maintaining effective coating thickness, achieving precision grinding and polishing of TBCs on aero-engine blade surfaces.