This paper proposes a non-destructive method for measuring the thickness of thermal barrier coatings on aero-engine turbine blades, based on three-dimensional point cloud data. Initially, a line laser sensor is employed to acquire 3D point cloud data from the blade surface. Subsequent processes, including point cloud clipping, filtering, and registration, are implemented to mitigate noise and minimize clamping errors. The local normal projection method is then utilized to compute the distance between the point clouds of the coated and polished blades and the point cloud of the initial substrate, thereby enabling the determination of coating thickness. Experimental results indicate that this method exhibits robust feasibility and precision in the non-destructive assessment of thermal barrier coating thickness. Compared to the metallographic microscope measurements, the average relative errors of the coating thickness calculations for three samples after spraying were 2.69%, 2.54%, and 2.07%, respectively, while the average relative errors after polishing were 2.79%, 2.66%, and 3.08%. Furthermore, this study analyzes the sources of systematic errors, including point cloud acquisition errors, registration errors, and thickness calculation errors, and presents corresponding improvement measures. Overall, the findings suggest that this method is suitable for a diverse range of coating materials and has the potential to significantly enhance the detection efficiency and accuracy of aero-engine turbine blades.