<p>The topology of a network plays a critical role in the consensus process of multi-agent systems. However, in large-scale multi-agent systems, the maintenance of redundant connections in the topology and low-impact information inputs during the state of agent updates will impede the convergence efficiency of the system. To solve this problem, we propose a multi-agent consensus algorithm based on topology optimized by pruning algorithm. First, the topology is pruned according to the priority of the agents with the aim of removing redundant connections and optimizing the topology. This effectively reduces unnecessary constraints and low-impact control inputs. Second, by analyzing the agent’s neighbor distribution, virtual nodes are created when specific conditions are met to accelerate the inward convergence of agents located at the edge of the system. Finally, we demonstrate the feasibility of the proposed method and provide a geometric proof of connectivity. A series of simulation experiments also demonstrate the effectiveness and superiority of our algorithm. The proposed algorithm provides an efficient solution for the consensus problem in large-scale multi-agent systems, significantly enhancing the scalability of such systems. It is particularly suitable for applications requiring fast convergence, such as rendezvous tasks.</p>