Distributed state estimate is one of the most fundamental problems for wireless sensor network. This paper addresses a type of distributed extended kalman filter that is extended from linear distributed kalman filter. Central extended kalman filter is an effective tool for nonlinear state filter of multisensor network. In this paper central extended kalman filter is decomposed into n micro extended kalman filters with inputs that are provided by consensus filters. When system process model and observation model are nonlinear, it is proved that distributed extended kalman filter can provide an identical state estimate of system state. Two target tracking examples are employed for simulation demonstration. All sensor nodes are able to take a nonlinear observation to moving target, dynamical cluster that is composed of several sensor nodes execute observation and error covariance matrix consensus filter. Each sensor in cluster obtain system estimate through distributed extended kalman filter. Simulation results show the proposed algorithm is effective for nonlinear distributed state estimate.