The unmanned aerial vehicles (UAVs) formation needs the frequent data-exchanging of individual’s state between units for monitoring and instruction uploading from the ground station, which inevitably occupies huge communication bandwidth. This paper presents the extended Kalman filter (EKF) design based on the event-triggered strategy to get UAVs greatly relieved of the communication burden with guaranteed accuracy. The event-triggered strategy firstly selects only the state measurements containing innovational information for the purpose of filtering. Due to the nonlinearity of UAV system, the EKF is further applied to make full use of the information from the prior event-trigger strategy so as to enhance the performance of estimation. The proposed algorithm is verified on the physical UAVs regarding the estimation quality and communication rate, demonstrating the robust dynamic performance with effectively reduced communication rate.