Cellular additively manufactured metallic structures for load-bearing scaffolds in the context of bone tissue engineering (BTE) have emerged as promising candidates. Due to many advantages in terms of morphology, stiffness, strength and permeability compared to conventional truss structures, lattices based on triply periodic minimal surfaces (TPMS) have recently attracted increasing interest for this purpose. In addition, the finite element method (FEM) has been proven to be suitable for accurately predicting the deformation behavior as well as the mechanical properties of geometric structures after appropriate parameter validation based on experimental data. Numerous publications have examined many individual aspects, but conceptual design procedures that consider at least the essential requirements for cortical and trabecular bone simultaneously are still rare. Therefore, this paper presents a numerical approach to first determine the actual admissible design spaces for a choice of TPMS based lattices with respect to key parameters and then weight them with respect to further benefit parameters. The admissible design spaces are limited by pore size, strut size and volume fraction, and the subsequent weighting is based on Young's modulus, cell size and surface area. Additively manufactured beta-Ti-42Nb with a strain stiffness of 60.5GPa is assumed as material. In total, the procedure considers twelve lattice types, consisting of six different TPMS, each as network solid and as sheet solid. The method is used for concrete prediction of suitable TPMS based lattices for cortical bone and trabecular bone. For cortical bone a lattice based on the Schwarz Primitive sheet solid with 67.572μm pore size, 0.5445 volume fraction and 18.758GPa Young's modulus shows to be the best choice. For trabecular bone a lattice based on the Schoen Gyroid network solid with 401.39μm pore size, 0.3 volume fraction and 4.6835GPa Young's modulus is the identified lattice. Finally, a model for a long bone scaffold is generated from these two lattices using functional grading methods in terms of volume fraction, cell size and TPMS type. In particular, the presented procedure allows an efficient estimation for a likely suitable biometric TPMS-based scaffolds. In addition to medical applications, however, the method can also be transferred to numerous other applications in mechanical, civil and electrical engineering.