In multi-layer complex networks, accessing node state information is often hindered by external factors, compounded by varying topologies across network layers. Building on prior research and drawing inspiration from real-world scenarios, this paper addresses the output coupling. Firstly, to achieve better control results and reduce control expenses, a dynamic event-triggered control strategy, leveraging local output information of nodes, is proposed. Then, synchronized states are utilized as virtual nodes to expand original network and reduce topological connectivity effectively. Since the coefficient matrix of output is no limited to being diagonal or positive definite, the model can be used to a wider range of situations. Furthermore, several sufficient conditions for achieving fixed-time synchronization are established using the Lyapunov method and inequality deflation technique, and Zeno behavior is excluded through rigorous theoretical analysis. Finally, to confirm the viability of the constructed controller and the accuracy of the theoretical conclusions, an example is given.