In this article, we tackle the problem of path following control of an underactuated autonomous surface vehicle in the presence of external disturbances and model parametric uncertainty. Using the conventional Lyapunov-based backstepping technique, we derive a nonlinear control law to determine a vectored actuation composed of a thrust direction and a thrust force that is bounded with respect to the position error. The model's linear and angular hydrodynamic damping terms are assumed to be known with uncertainty, while the external disturbances are fully unknown. The errors stemming from both the model parametric uncertainty and the exogenous disturbances are compensated by a couple of bounded integral actions which, when embedded into the control law, help rendering the origin of the error system a global attractor. The proposed strategy is validated by a set of simulation results, with its efficacy and robustness further demonstrated through a set of experimental trials using a fully autonomous instrumented surface craft.