<p>This paper presents an adaptive fault-tolerant trajectory-tracking control scheme for underactuated surface vessels under quantized-state feedback. Motivated by bandwidth-limited marine environments, all system states are quantized before feedback transmission, which introduces non-smoothness into the backstepping design. To avoid differentiating the resulting virtual control laws, a standard first-order command filter is employed. Neural networks are used to approximate uncertain nonlinearities and external disturbances. By integrating coordinate transformation, command-filtered backstepping, and adaptive fault accommodation, the proposed framework addresses the coupled effects of underactuation, state quantization, and actuator faults. Theoretical analysis shows that the quantization effects and all closed-loop signals are bounded without requiring a global Lipschitz condition. Simulation results on underactuated surface vessels demonstrate the effectiveness of the proposed method.</p>