This paper addresses the challenge of achieving precise trajectory tracking in the vertical plane for autonomous underwater gliders amidst model uncertainties and external disturbances. With the aim of enhancing control robustness, we propose a novel approach that integrates neural networks, sliding mode control, and backstepping techniques. This multifaceted control system is designed to effectively regulate the glider’s motion dynamics, considering the inherent constraints imposed by the ballast and moving masses. Through comprehensive Lyapunov stability analysis, we establish the boundedness of all signals and the eventual convergence of tracking errors towards zero, ensuring reliable performance even in the presence of unpredictable environmental factors. Notably, our proposed controller offers practical feasibility by eliminating the need for precise knowledge of glider parameters or external disturbances, thereby simplifying implementation in real-world scenarios. The effectiveness and robustness of our approach are corroborated through simulation studies, highlighting its potential for enhancing the operational capabilities of autonomous underwater gliders. This research contributes to the advancement of underwater robotics, offering a viable solution for achieving precise trajectory tracking in challenging marine environments.