In the context of event-triggered multi-agent attitude synchronization control, considering the presence of the dynamically changing leader reference attitude which is not globally known, this paper first proposes a novel event-triggered state estimator. This estimator is utilized to provide reference attitude information for the follower agents. Subsequently, combining by the state estimator, a control algorithm for attitude synchronization is designed. Both state estimator and attitude synchronization control algorithm employ a unified event-triggering function which is designed in this paper. Finally, the overall closed-loop stability is demonstrated. The algorithm ensures the asymptotic stability while reducing the communication frequency among formation members, effectively reducing communication energy consumption. Moreover, when the bandwidth of the multi-agent formation is limited, the algorithm can reduce the communication burden. Furthermore, numerical simulations validate the efficacy of the algorithm. Additionally, this paper demonstrates that under the operation of this algorithm, there is an absence of Zeno behavior within the system.