Air-ground collaborative multi-unmanned systems represented by unmanned aerial vehicle (UAV) and unmanned ground vehicle (UGV) have significant applications in collaborative mission execution. However, grouping and tracking between leaders and followers is a challenging problem in the field of multi-intelligence grouping and formation. This paper proposes a leader autonomy decision-making method based on weighting factors to address this issue. A three-layer control architecture comprises virtual leaders, group leaders, and followers for fixed-time grouping formation control. Firstly, weight factors are introduced based on the dynamic potential field method, allowing grouping decisions for followers to be made by group leaders according to weight information. Thus, enabling different groups to undertake distinct formation tasks. Subsequently, a grouping formation controller has been developed based on consensus theory and fixed-time theory, facilitating the convergence of the system to achieve the desired formation within a finite time frame. Finally, the results of the physical simulations performed in Gazebo verify the accuracy of autonomous decision-making and the ability to form formations in a finite time, with fast convergence and robustness.