Aiming at the problem that it is difficult for cluster UAVs to function with the combination of traditional GPS and inertial navigation when satellite signals are obstructed, a cooperative navigation algorithm for cluster UAVs is designed, in which UAVs equipped with low-precision inertial devices acquire the position information of UAVs equipped with high-precision inertial devices within the cluster through a data chain, and the position, velocity, and attitude information outputted from their own low-precision inertial devices is corrected by the trace-free Kalman filter to reduce the navigation error. And attitude information to reduce the navigation error. After 1100 s simulation, the latitude and longitude errors of the low-precision UAV converge to 1 × 10−6 degrees, which is about 0.1 m, the northward and eastward velocity errors are reduced to 0.01 m/s, and the pitch, heading, and roll angle errors are less than 10% of the initial alignment error, which improves its navigation accuracy.