Abstract: A robust terminal sliding mode controller based on adaptive filtering backstepping was designed for the ship course tracking problem. The closed-loop system consists of the 2-nd order nonlinear Nomoto's autopilot model and the rudder response dynamics. The unknown terms, which were approximated online by using the fuzzy logic system, include the model uncertainties and the external disturbances. The filtering backstepping method avoids computing the derivatives of the virtual control by introducing a 2-nd order low pass filter, and consequently simplifies the backstepping implementation. Moreover, the low pass property of the filter decreases the effects of the measurement noise. Adaptive filtering backstepping was then used to design the terminal sliding surface, and finally a variable-structure controller was derived, which can guarantee the finite-time convergence of the tracking errors. According to the Lyapunov approach, all the error signals in the closed-loop system are proved to be uniformly ultimately bounded. Simulation results demonstrate the effectiveness of the proposed method.