Abstract: The unstructured-grid finite volume method based on implicit time stepping is one of the most commonly used numerical schemes in computational fluid dynamics (CFD). Due to the irregularity of unstructured grid cell ordering and to the frequent need to access neighboring cell information during a computation, the quality of the grid cell ordering directly impacts the computational efficiency. The widely used Reverse Cuthill-McKee (RCM) reordering algorithm can significantly reduce the bandwidth of implicit operators and decrease the number of implicit iterations. Meanwhile, the reordering algorithm based on the Hilbert space-filling curve has been demonstrated to effectively improve data locality. To develop a reordering algorithm that balances both implicit iterative updates and data locality, the Morton and Peano space-filling curves are extended to unstructured grid reordering and compared with the RCM algorithm and the Hilbert-based reordering method. The study finds that the Morton space-filling curve-based reordering algorithm possesses characteristics similar to those of the RCM algorithm while retaining the inherent property of space-filling curves to enhance data locality. Test research results show that the Morton-based reordering algorithm reduces the execution time by 40% ~ 50% compared to the initial grid ordering, and achieves approximately 10% greater reduction in the execution time than those of both the RCM and Hilbert-based reordering algorithms.