This paper proposes an event-triggered communication (ETC) mechanism to solve the nonlinear multi-agent system synchronization (coordination) problem for networks whose communication topologies are described by weight-balanced, strongly connected digraphs. To reduce communications among vehicles, each agent uses the dynamic model of its neighbors to estimate the states of its neighboring agents. In order to decide when each agent will broadcast its state, two types of event triggering functions are proposed: one that is purely time-dependent, and another, called state-dependent, that depends on the estimated states of the neighbors and the local information of the agent itself. We show that with the proposed ETC mechanism, coordination of a multi agent system is achieved. The proposed strategy also guarantees that the minimum of the inter-event times for every agent is strictly positive, hence Zeno behavior is excluded. The results are illustrated via simulations.