Abstract: As an emerging lightweight multifunctional material, graded lattice structures have gained extensive applications in various high-technology fields, including aerospace, medical devices, and automotive industries, owing to their exceptional mechanical properties, such as high specific strength and specific stiffness. However, the unit cell distributions of most existing lattice structures are predominantly designed based on uniform distribution assumptions, consequently limiting their ability to fully exploit mechanical performance under diverse boundary and loading conditions. To address these limitations, this paper proposes a graded hierarchical lattice optimization design method based on intelligent optimization algorithms. Initially, an explicit topology description model for lattice unit cells is constructed using level-set functions, with shape interpolation techniques introduced to ensure geometric continuity between hierarchical layers. Subsequently, a functional relationship between the relative density of unit cells and their equivalent mechanical properties is established through a Kriging surrogate model. An optimization model for graded hierarchical lattice structures is then formulated, aiming to minimize overall structural compliance while constrained by structural volume fraction and boundary conditions. The Dung Beetle Optimizer algorithm is employed to solve this model efficiently. Concurrently, shape interpolation techniques refine geometric continuity in the final optimized structure. Validation via numerical examples and simulation experiments confirms the method’s effectiveness. The load-bearing performance of the optimized graded hierarchical lattice structure is compared with uniform lattices and topology-optimized graded lattices under identical conditions. Results demonstrate a compliance reduction of 70.72% and 10.27%, respectively, proving superior mechanical performance. This approach provides a novel design paradigm for high-load-bearing lightweight lattice structures.