The increasing demand for lightweight yet mechanically robust components has promoted significant interest in lattice structures. Owing to their superior strength-to-weight ratio, compared to traditional bulk materials, the lattice architectures are well suited for structural applications, and they can be readily produced by additive manufacturing. This study investigates the elastic-plastic behaviour of Polyamide 12 (PA12) lattice structures, manufactured with the Multi Jet Fusion (MJF) technique, combining experimental, numerical, and theoretical approaches in order to establish a comprehensive framework for their mechanical homogenization. Tensile tests on bulk specimens were initially conducted to characterize the constitutive behaviour of the material. The resulting elastic-plastic model was then applied in the homogenization process, which relied on periodic boundary conditions imposed on a representative volume element (RVE), and this was achieved by means of Finite Element simulations, with solid elements, and dedicating particular attention to the symmetry conditions. Since the hardening response of the RVE depends on the loading direction, a simplified model is proposed to capture the anisotropic plastic behaviour. This approach combines Hill’s yield criterion with the Levy-Mises plastic flow rule, and incorporates direction-dependent hardening, enabling efficient prediction of lattice performance under complex loading scenarios. Homogenized elastic-plastic properties were then evaluated for varying unit-cell aspect ratios (AR), leading to mathematical expressions that describe the dependence of the mechanical properties on the AR. The homogenization methodology was validated both numerically and experimentally using two geometries of optimized graded strut-based lattice specimens. Numerical and experimental tensile tests on these lattices were compared with simulations performed on homogenized models, demonstrating agreement and then confirming the predictive capability of the proposed framework. Overall, the findings highlight the potential of this methodology for structural optimization and mechanical performance prediction in components demanding lightweight and high strength. In particular, a biomedical application is introduced as a target of the proposed material and manufacturing framework, combined with the proposed modelling methodology.