In the domain of autonomous air combat decision-making, there exists a significant disparity between offensive and defensive maneuvers. Utilizing a single intelligent decision-making framework fails to effectively address both aspects simultaneously. This paper proposes a hierarchical autonomous air combat decision-making framework, which is based on the tactical pursuit point method at its core. This framework separately designs autonomous offensive and defensive strategies at the lower layer, while completing the autonomous evaluation and selection of battlefield attack and defense phases at the higher layer. Building upon the proposed solution, this paper further decomposes the functionalities of each layer and provides design examples based on the tactical pursuit point method. The proposed framework exhibits versatility and can be applied to other intelligent decision-making methodologies.