This paper presents a robust adaptive nonlinear proportional--integral (ANPI) scheme to control the speed of a direct-current (DC) motor. Unlike proportional--integral--derivative (PID) controllers, PI controllers have a simpler structure and they deliver effective control effort. However, due to inadequate controller gains, they are often unable to simultaneously improve the transient as well as the steady-state response of the system. A nonlinear PI (NPI) controller alleviates these issues and delivers a good response. In this research, the proportional and integral gains of the NPI controller are dynamically modulated via a nonlinear sigmoidal function (SiF) of the error dynamics of the motor's angular speed. The variation rates of these functions are manually tuned via trial-and-error method. These rates are also dynamically updated via an extended Kalman filter (EKF)-based adaptation mechanism. The performances of a linear PI controller, an NPI controller having fixed variation rates, and an NPI controller equipped with the EKF-based self-regulated SiFs are tested and compared in real time. The experimental results are analyzed to validate the effectiveness of the proposed ANPI controller in optimizing DC motor speed control.