Unmanned network-flying vehicles (UNFVs), regarded as special remote-sensing platforms, are affected by attitude maneuvering, gravity, climatic change, and atmospheric turbulence. To improve the positioning precision of the UNFVs and accomplish autonomous navigation in the case of communication and/or tracking failure, a tracking positioning algorithm with joint space-time constraints (PAST) is proposed. In particular, the adaptive truncated normal distribution model is adopted to describe the distribution of the vehicle maneuvering acceleration with reduced delay effect of the vehicle tracking. Then, according to the spatial correlation of the UNFVs in adjacent moment, a spatial adjustment datum is combined with a time-domain strong tracking filter by Lagrange method of multipliers. As such, the rank deficiency in the measurement equation is resolved while the spatial correlation of the UNFV tracking enhanced. Furthermore, for the autonomous navigation of the UNFVs which depends little on the Earth- or satellite-based tracking, telemetry, and command, we utilize an ldquoanchorrdquo technology to further improve the PAST algorithm. Monte Carlo simulation results show that, compared with existing positioning methods, the proposed algorithm may continuously track sudden motion, rapidly respond to network communication failure, and greatly increase positioning precision for slowly moving vehicles.