<p>This paper proposes a multi-material topology optimization method for designing porous load-bearing structures with convective cooling capabilities. To efficiently account for convective cooling during optimization, a porous-feature-based convective heat transfer coefficient calibration method (PCHCM) is first developed. Using a design-variable-gradient-based representation of interfacial intensity together with case-specific equivalent parameters, the PCHCM establishes a two-dimensional equivalent thermal model for short three-dimensional channel configurations, thereby avoiding repeated fluid analyses during optimization. Its accuracy and applicability are verified through finite element analyses on representative structures under different flow conditions and geometric variations. Based on the calibrated method, topology optimization is performed on two-dimensional cross-sections equivalent to three-dimensional load-bearing cooling structures, where fluid regions act as coolant channels and solid regions consist of multiple materials with different thermal and mechanical functions. Thermal and mechanical performances are combined through a normalized weighted-sum formulation, and a maximum-length projection filter is introduced to control pore size. Numerical examples show that the proposed method can generate porous structures with balanced distributions of coolant channels and functional materials. The results further demonstrate satisfactory thermal accuracy, significantly improved computational efficiency, and enhanced robustness under local failure scenarios, while maintaining a favorable balance between cooling effectiveness and load-bearing capacity.</p>