<p>This paper provides a comprehensive review of homogenization procedures for lattice structures, explicitly distinguishing between homogenization techniques and homogenization approaches. Homogenization techniques, including beam theory, strain-energy equivalence, micropolar theory, asymptotic homogenization, and FE² technique, form the basis of formulations for determining effective mechanical properties by simplifying complex microscale geometries into equivalent macroscale descriptions. The analytical, numerical, and data-driven homogenization approaches are selected to compute effective properties, with a focus on computational efficiency, accuracy, and practical applicability. The review systematically assesses the strengths, limitations, and practical trade-offs associated with each technique and approach, explicitly considering computational costs, robustness against manufacturing-induced imperfections, and sensitivity to geometric variations. Finally, the paper identifies significant research gaps. It proposes future research directions, particularly advocating for the integration of advanced methodologies such as machine learning (ML) with traditional computational homogenization frameworks to effectively tackle challenges related to complex lattice geometries and nonlinear structural behavior, as well as full-cycle multiscale techniques for the reliable prediction of failure loads of the lattice structures.</p>