The unpredictable nature of the marine environment , combined with nonlinear dynamics and parameter uncertainty of underwater vehicles makes the control system design for such vehicles a challenging task. Based on these issues, hybridising robustness and adaptation in the control system could result in more successful marine missions. This work proposes a robust adaptive control (RAC) scheme for trajectory tracking of an autonomous underwater vehicle. The proposed RAC scheme has been developed by exploiting the advantages of a robust sliding mode controller (SMC) and an adaptation law. Lyapunov arguments are proposed to prove the exponential stability and finite-time convergence of the resulting closed-loop dynamics tracking error to an invariant set, S (very close to zero). Scenarios-based real-time experiments are conducted with the Leonard ROV prototype to demonstrate the effectiveness of the proposed RAC approach. The control design performance indices (root mean square error RMSE, integral absolute error IAE and integral square error ISE) and a comparative analysis with a recent control scheme from the literature confirm the interest of the proposed RAC scheme for marine applications.