Current research efforts have yet to comprehensively compare the thermo-mechanical performance disparities between layered and power-law graded graphene-reinforced composites, particularly regarding structural optimization under operational constraints. To address this gap, this investigation establishes a model methodology for composite analysis, synergistically integrating the Halpin-Tsai micromechanical constitutive modeling framework with Maxwell-Eucken multiphase homogenization principles. A nonlinear finite element framework was established based on the simplified first-order shear deformation theory (S-FSDT), and the numerical solution of the bending response of the plate is obtained by Newton-Raphson iteration. The proposed models were validated against existing literature. A systematic study was conducted on the thermodynamic coupling properties of layered and power-law graded graphene distributions with various pore types. Parametric comparisons showed that symmetric surface-enriched distributions achieved optimal performance in all configurations. The power-law graded type demonstrated superior reinforcement efficacy over the layered design. This study reveals the synergy between pore distribution and graphene gradient design, offering theoretical support for optimizing lightweight high-stiffness composites.