Buoyancy control of underwater robot vehicles (URVs) has been a challenge due to the high current load of the ocean. In this work, the shape of a spherical URV is presented and a mathematical modeling formulation are given of a variable ballast tank system and the forces affecting its controller. The URV system is simulated under different operational conditions using a fuzzy logic controller and a neural network controller. The results show that both controllers give a good performance but in terms of fast response the fuzzy logic controller produces a better result compared to the neural network controller. In the experimental setup, the prototype is tested and verified for downward and upward trajectories and an experimental comparison is carried out. These techniques have wide applications in underwater activities such as oil exploration, pipeline maintenance and research activities. © 2017, UK Simulation Society. All rights reserved.