Spectrally selective absorbers garner significant attention across diverse domains owing to their pivotal roles in electromagnetic stealth technologies, solar-thermal photovoltaics, and related applications. However, enhancing the absorption properties frequently necessitates the augmentation of the metamaterial patterned layer complexity. This introduces a paradox in application, where the increased intricacy of structural patterning adversely intersects with fabrication processes, thereby exacerbating the practical applicability challenges due to manufacturing constraints. Therefore, this study leverages a design methodology that combines artificial intelligence (AI) with finite element simulation. This approach propels the realization of broadband selective absorption based on a simple biomimetic metamaterial structure, achieving broadband absorption without increasing structural complexity or reducing fabrication efficiency. The spectrally selective absorbing metamaterial designed with AI achieves broadband absorption unaffected by polarization in the 5-8 mu m range. With electromagnetic waves impinging perpendicularly, the average absorptance exceeds 0.9, proving valuable for radiation cooling compatible with infrared stealth. Furthermore, the design method elucidated in this study exhibits remarkable robustness and transferability, significantly improving the design efficiency of complex spectral metamaterials. This innovative approach heralds a design paradigm shift, facilitating the creation of stealth-compatible and other advanced multiband spectrally selective absorbing materials.