Aimed at problem that unmanned combat air vehicle exists in autonomous air combat tactical decision making, this paper presents an approach of case-based reasoning (CBR) and rule-based reasoning (RBR) and a method for describing autonomous air combat tactical decision making cases based on frame structure. The research shows that the improved nearest neighbor approach can improve the effectiveness and accuracy of case-based reasoning by introducing the calculation method of structural similarity and the theory of cloud model. Finally, the result of simulation shows that any appropriate solution can be selected from the case library.