This paper considers underwater wireless sensor networks (UWSNs) for surveillance and monitoring. Sensors are distributed in several key sections along the seafloor to record the surrounding environment, for example, monitoring oil pipelines and submarine volcanoes. Due to the need for timely data reporting and the fact that underwater communications suffer from a significant signal attenuation, homogeneous autonomous underwater vehicles (AUVs) are sent to retrieve information from the sensors, and periodically surface to report the collected data to the sink. In this paper, considering the huge energy consumption of surfacing and diving, our objective is to determine a trajectory schedule for the AUVs so that the total amount of surfacing for all the AUVs are minimized, and the data is reported to sink within the deadline. We first investigate the influence of different movement directions of AUVs, and provide the optimal solution to minimize the amount of surfacing for multiple AUVs within the same sensor section. Then, we propose a greedy detouring scheme to collaboratively schedule the AUVs in adjacent sensor sections. Extensive experiments show that our trajectory scheduling improves performance significantly.