In some servo applications, restricted by the measurement noise and the stiffness of the mechanical system, the bandwidth of the disturbance observer cannot be high enough to realize a good robust control to parameters uncertainties, it is essential to identify the mechanical parameters. In the existing mechanical parameter identification methods, the viscous and Coulomb friction torque coefficients are usually identified using two different constant speeds, and the moment of inertia can be identified with various methods. However, no one method can identify these three parameters at the same time using only one speed. The contribution of this article is to propose a method that can identify the three mechanical parameters simultaneously using only one speed. A low-pass filter is used to suppress the measurement noise in the speed feedback, thus parameters are identified with high precision. A unidirectional sine-wave speed is used to decouple the viscous friction torque and the Coulomb friction torque, two speeds with different amplitudes and the detection of the speed amplitude are no more required. Adaptive extended state observer (ESO) is designed based on the Lyapunov's stability theory, can be easily changed to the normal ESO used in active disturbance rejection control. The proposed method is easy to understand and implement. The effectiveness of the method is verified by both the simulation and experimental results.