The development of ubiquitous intelligence has put forward new requirements for the coverage density of system sensing and communication functions. Through the collaborative work of air and ground nodes and the optimization of node coverage in the air-ground integrated sensing and communication (ISAC) network, real-time monitoring of a specific geographical area can be achieved. However, the dynamic characteristics and energy constraints of nodes make it a major challenge to adaptively adjust the deployment and working status of nodes to achieve optimal coverage. To this end, this paper proposes an adaptive multi-dimensional optimization method for air-ground node distribution and coverage, which is formalized as a multi-objective optimization problem. Multiagents of air-ground multi-nodes are constructed to collaboratively optimize the global strategy. By introducing more complex reward functions, the agents can comprehensively consider multiple core factors such as energy consumption, coverage hole repair, and node connectivity while optimizing network coverage. Furthermore, in order to improve performance and efficiency, this paper constructs a generative AI model. On the one hand, by generating simulated environments and samples, it learns how to dynamically adjust weights when weighing multiple objectives; on the other hand, it is used to generate candidate strategies to find the best balance between multiple objectives. Experimental results show that the algorithm shows significant advantages in improving network coverage, node connectivity, and energy utilization efficiency, providing effective support for the actual deployment of ISAC networks.