Abstract: To address the issues of traditional manual positioning dependency and potential safety hazards in the riser cutting process of railway wheels, a cutting path planning method based on 3D point cloud is proposed. Firstly, the 3D point cloud of the railway wheel riser scene is acquired using an industrial 3D camera, followed by point cloud preprocessing through adaptive height threshold segmentation, circular region constraint, and statistical filtering. Secondly, the α-shape algorithm is employed to extract the original boundary points. A secondary screening method combining "vector interpolation + density peak selection" is proposed for the original boundary points, and the workpiece center positioning is achieved through linear least squares solution and nonlinear least squares geometric optimization. Thirdly, the ear hole center is optimized with reference to the pixel center of the ear hole. Finally, a vector-based positioning method for preheating points and cutting points is proposed to obtain the complete cutting path. The experimental results demonstrate that the maximum cutting error of the proposed system is 0.32 mm, which verifies the effectiveness and industrial feasibility of the method. The proposed approach meets the precision requirements of industrial production.