In this paper, we investigate a mobicast, also called a mobile geocast, problem in three-dimensional (3-D) underwater sensor networks (USNs), which aims to overcome the hole problem and minimizes the energy consumption of the sensor nodes while maximizing the data collection. In this paper, all underwater sensor nodes are randomly distributed in a 3-D underwater environment in the sea to form a 3-D USN. Considered a mobile sink or an autonomous underwater vehicle (AUV), all possible sensor nodes near the AUV form a 3-D geographic zone called a 3-D zone of reference (3-D ZOR). The AUV travels a user-defined route and continuously collects data from sensor nodes within a series of 3-D ZORs at different times. The main problem is how to efficiently collect data from sensor nodes within a 3-D ZOR while those sensor nodes are usually in sleep mode for a long period. The routing protocol relies on two phases: the first phase consists of collecting data within a 3-D ZOR, and the second phase consists of waking up those sensor nodes in the next 3-D ZOR to be queried while trying to avoid topology holes. To save power, only sensor nodes in a 3-D ZOR are notified to enter the active mode in order to deliver sensed results to the AUV. The specific characteristics of USNs, including low communication bandwidth, large propagation delay, and ocean current, are significantly different from wireless sensor networks. To consider the characteristics of USNs, a new mobicast routing protocol is developed in 3-D USNs. The key design challenge is to develop a power-saving mobicast protocol in 3-D USNs to overcome the unpredictable 3-D hole problem. To solve the hole problem, an “apple slice” technique is used to build multiple segments to surround the hole and to assure routing path continuity. Finally, performance analysis is derived, and simulation results illustrate the performance improvement in successful delivery rate, power consumption, and message overhead.