This paper addresses the yaw control problem of underwater robots operating under strong disturbances and nonlinear uncertainties by proposing a fuzzy logic-enhanced active disturbance rejection control (F-ADRC) method with an improved nonlinear function. Active disturbance rejection control (ADRC) is utilized to improve disturbance rejection performance in complex environments. A three-segment nonlinear function is incorporated to enhance control accuracy and sensitivity in small-error regions. To reduce the reliance on empirical parameter tuning and improve adaptability, a fuzzy controller is designed to dynamically adjust both the nonlinear state error feedback (NLSEF) gains and the extended state observer (ESO) bandwidth. Simulation results based on a horizontal-plane underwater robot model demonstrate that the proposed F-ADRC method achieves faster response, smaller steady-state error, and superior disturbance rejection compared to conventional PID and standard ADRC approaches.