To improve the effectiveness of multiple unmanned aerial vehicles(UAVs) cooperative target tracking, a multi-UAVs cooperative target tracking algorithm based on co-optimization of communication and observation is presented. In the packet-erasure channel model, the probability of the successful information transmission is modeled as a function of the signal-to-noise ratio. The extended information filter is build for the estimation fusion and prediction of target states. The Fisher information matrix is applied to express the amount of information obtained from the measurements. The basis for co-optimization of communication and observation is the derivation of a scalar quantity which is a function of the expected information gain and the probability of the successful information transmission. The co-optimization of communication and observation is reflected in the tradeoffs between communication performance and sensing utility to improve overall information content delivered to the base station. Informative trajectories that maximize this metric are generated in the frame of receding horizon optimizing. Simulations demonstrate that the proposed algorithm can effectively improve the estimation performance relative to the approach that does not consider communication reliability.