This article proposes a solution for complex dynamic multi-agent systems based on the current research status. By analyzing traditional leaderless multi-agent systems, they are divided into two categories, and the necessary and sufficient conditions for the existence of consensus protocols are given respectively. The shortcomings of this type of model are also analyzed. Considering the shortcomings of traditional models, a new dynamic model with a leader is proposed, which can flexibly transform the target state and facilitate practical operation. This article attributes whether the model can achieve state consistency to the existence of solutions to a matrix equation system. Based on the relevant theory of dynamic decentralized feedback in large-scale systems, a sufficient condition for the existence of solutions to matrix equation systems is given.