This paper presents a robust neural control algorithm with a dynamic fault compensation mechanism for maritime dynamic positioning systems. The proposed dynamic fault model integrates the effects of time-varying disturbances and uncertain actuator wear and degradation. The algorithm employs a radial basis function neural networks to approximate the uncertainties in the model and utilizes robust damping techniques to compress the weights and errors of the neural network. By estimating the gain and degradation coefficients and establishing an adaptive law for online updating, the algorithm compensates for uncertainties in the servo system’s gain and the dynamic variations in the degradation coefficients within the fault model. An extended configuration matrix is introduced for thrust allocation in the actuators, converting the actual control inputs into the actuator’s pitch ratio and azimuth angle. Finally, the Lyapunov stability theory is applied to prove the boundedness of all signals in the closed-loop system. Simulation results demonstrate the effectiveness of the proposed algorithm.