Spinodoid mechanical metamaterial has received increasing interests since last decade because of its tunable stochastic design and defect insensitive. However, the mechanical response of spinodoid mechanical metamaterials is strongly correlated with generated stochastic designs, which might offer very poor mechanical performance and very large variation of mechanical properties (e.g., stiffness and anisotropy), thereby limiting wider applications in industry. To tackle this designing challenge, we exploit the concept of hierarchical design, which has been proved a successful solution in Nature to generative tunable enhanced mechanical behaviors of materials, to improve the mechanical response of spinodoid mechanical metamaterials. Moreover, to effectively explore the huge design space offered by hierarchical spinodoid mechanical metamaterials, artificial intelligence (AI) based approach will also be exploited.