<p>For the cooperative control problem of multi-agent systems subject to complex unknown disturbances and communication resource constraints, this paper proposes a hierarchical disturbance observer-based event-triggered compound control strategy integrated with adaptive frequency-domain decomposition. A hierarchical disturbance observer is constructed by combining a fast response layer with an accurate estimation layer. Compared with the classical extended state observer (ESO) and active disturbance rejection control (ADRC), the proposed hierarchical disturbance observer (DOB) achieves online frequency-domain separation of complex disturbances, which solves the contradiction between rapidity and accuracy in ESO/ADRC and improves the adaptability to multi-frequency composite disturbances. By leveraging adaptive frequency-domain decomposition, disturbances are analyzed and separated online in the frequency domain, thereby significantly enhancing both the accuracy and speed of disturbance estimation. Within a backstepping control framework, a radial basis function neural network is employed to approximate the system’s nonlinear dynamics, enabling the design of an adaptive controller with guaranteed stability. Furthermore, an adaptive event-triggering mechanism based on dynamic state regulation is developed to intelligently adjust the communication threshold, thus achieving efficient utilization of limited communication resources. Based on Lyapunov stability theory, it is rigorously proven that all signals in the closed-loop system are semi-globally uniformly ultimately bounded. Simulation results demonstrate that the proposed approach achieves substantial improvements in both unknown disturbance estimation accuracy and tracking control performance, while significantly reducing communication overhead, thereby validating the effectiveness and superiority of the proposed theoretical framework.</p>