In the realm of designing thermoelastic structures, particularly within aerospace and broader engineering fields, complex challenges arise from the diverse behaviors of heterogeneous materials and the intricate requirements of stress-based systems in coupled thermomechanical considerations. This study addresses these challenges by proposing a comprehensive methodology with two main contributions: (i) the development of an effective unified solution for stress-driven designs that tackle heat-sinking problems by accommodating a wide range of materials, from homogeneous to heterogeneous and (ii) the introduction of a robust stress-based optimization approach for multi-material problems within coupled thermomechanical systems. The methodology employs the well-established P-norm approach to consolidate heterogeneous stresses into a unified global metric. Additionally, it integrates effective material interpolation with the generalized Solid Isotropic Material with Penalization (SIMP) framework, resulting in comprehensive multi-material models that include the constitutive matrix, thermal conductivity, thermal stress coefficient, and thermoelastic stress distributions. To further enhance adaptability and flexibility, the methodology incorporates a polygonal discretization technique. Detailed sensitivity analyses, using the adjoint variable technique, are conducted to improve computational efficiency in gradient-based mathematical programming algorithms. The efficiency, robustness, and practicality of the proposed methodology are validated through numerical examples, demonstrating its effectiveness and reliability in real-world applications.