In this paper, a longitudinal motion control system for a fully submerged hydrofoil is studied. A feed forward multilayered neural network is used for craft forward dynamics. The neural network is trained in a quasi-online regime in order to cope with changing dynamics of the craft. A feedback linearization controller for attitude control is designed. The performance of closed loop system is verified through simulation. The closed loop system stability is proved based on Lyapunov function theory. The simulation results show that the proposed control system is able to keep the craft in predefined height and serves the consideration in designing height control system for real crafts.