This study proposes a navigation method of multiple autonomous underwater vehicles (AUVs) where low-cost AUVs (child AUVs) estimate their self-states including their position, yaw angle, and horizontal moving velocity accurately thanks to a single high-performance AUV (parent AUV) without any external support. Since the child AUVs have only simple navigational sensors, positioning performance is not enough. The parent AUV, on the other hand, can conduct accurate self-state estimation using high-grade navigational sensors without any external support. The parent AUV supports the child AUVs through all processes of deployment, surveying, recovery, and surfacing. This study focuses on the accurate positioning method for the phase of the surveying. The child AUVs have acoustic devices for positioning and communication or cameras; hence, they build a parent-centered wireless positioning and communication network and can also estimate their self-states as accurately as the parent AUV, thanks to the positioning relative to the parent AUV. Field experiments with AUVs were conducted in a deep sea (hydrothermal vent field) and a shallow area. We confirmed that long-range navigation could be achieved while keeping an accuracy that is 5–10 times greater than that achieved by the child AUV alone based on state estimation using sensor data from the experiments in postprocessing. The performance of the proposed system does not depend on the child AUV's performance. We also verified that the proposed method has a certain scalability and a high-level performance in terms of accuracy, efficiency, and cost by a comparison with other existing navigation methods.