In a wild battlefield environment lacking infrastructure, there are many heterogeneous intelligent devices with different resources and dynamic movement. Moreover, some of these intelligent devices (also called agents) have limited computing, storage, and other resources, so a single device may not be able to meet the resource requirements for computing-intensive, delay-sensitive, complex process tasks, etc. This requires multiple agents to collaborate to complete these tasks and provide quality assurance for task execution. In response to this situation, this paper proposes a task scheduling strategy based on computing-aware and multi-agent collaborative services (CAMC), which aims to achieve the deterministic delay required for task completion through device collaborative services. To this end, we first designed a distributed computing-aware mechanism to obtain the resource status of accessory devices in a timely manner. Then, we optimize the task scheduling process by comprehensively considering the resource factors such as computation, storage, and transmission of each device, so that multiple devices can collaborate on services for the processing of tasks. Experimental results show that the proposed strategy can effectively reduce the task completion time, and the task completion rate has significant advantages compared with other representative algorithms.