<p>Autonomous underwater vehicles (AUVs) require robust and precise trajectory tracking to operate effectively in complex and uncertain underwater environments. This study proposes a neural-network-based backstepping control strategy designed for full six-degree-of-freedom (6-DOF) AUV dynamics with explicit integration of a thruster distribution matrix (T-thruster matrix). The T-thruster matrix maps individual thruster forces to resultant torques and forces, enabling more responsive and accurate motion control. The controller synergistically combines model-based backstepping with a radial basis function (RBF) neural network to approximate unknown nonlinearities, while an adaptive compensation term mitigates ocean current disturbances and approximation errors. Stability of the closed-loop system is rigorously guaranteed through Lyapunov theory and Barbalat’s lemma. Numerical simulations demonstrate that the proposed method achieves superior trajectory tracking, reduced errors, and enhanced robustness compared with existing approaches. The results underscore the practical importance of incorporating 6-DOF modeling, thruster integration, and disturbance handling for reliable AUV operation in dynamic underwater conditions.</p>