With the growing demand for cross-domain collaborative applications of unmanned systems, such as collaborative patrolling tasks among different unmanned platforms like drones, unmanned vehicles, and unmanned boats, multi-platform cross-domain communication has become a research hotspot. To address issues such as differing communication characteristics between platforms, unstable link communication quality, and network congestion, this paper proposes a dual-layer cross-domain communication framework based on transmission service requirements. The framework consists of a link quality assessment layer and a routing decision control layer. The link evaluation layer dynamically assesses the link communication quality by calculating the communication interruption probability of different transmission services, based on transmission characteristics in various communication scenarios, such as air-to-air, air-to-ground, and air-to-sea. The routing decision control layer perceives link communication quality, link lifetime, and network congestion conditions, using link coefficients and link congestion coefficients as key metrics for route selection. This enables dynamic switching among single-path, backup multi-path, and parallel multi-path transmission modes, thereby enhancing the network’s robustness and communication service quality. Finally, simulation results validate the effectiveness of the proposed method, providing new solutions and practical insights for future UAV ad hoc networks and data transmission in high-dynamic environments of cross-domain communication networks.