Titanium alloys, with their excellent corrosion resistance, light weight and high strength, have become a core material in the aerospace industry. With the rapid development of modern equipment manufacturing technology, the manufacturing technology of titanium alloy castings has also reached a new height. Not only is the complexity significantly improved, but the scale of the castings also tends to increase. The trend of large-scale, thin-walled and complex titanium alloy castings undoubtedly brings challenges to the numerical simulation of the casting process. The mesh is the computational domain of the numerical simulation process, so the primary issue is how to efficiently generate high-quality meshes. To address this problem, this paper proposes an algorithm aimed at efficiently generating hexahedral uniform meshes suitable for finite difference methods. The algorithm combines the octree data structure, which has the advantage of spatial division, and the classical ray piercing algorithm. The octree data structure is first utilized to perform fine spatial delineation of the 3D model of titanium alloy casting, and an accurate surface mesh is quickly generated through a mapping relationship with a finite difference mesh. Then, the ray piercing algorithm is used to accurately construct the internal mesh based on the surface mesh and the intersection information between the rays and the interior of the casting. After testing and verification, the algorithm can efficiently and accurately generate high-quality finite difference hexahedral uniform meshes, which provides an accurate and reliable computational domain for subsequent numerical calculations.