In the current age of automation, robots have been designed to interact and work with humans. Humanoid robots are mostly used for this social interaction. To interact in a social environment, robots need high decoupling of a camera sensor from the head skeleton to point, track and recognize the stimuli. This high decoupling of a camera sensor from the head skeleton is easily achieved by using the decoupling property of a gimbal system. To further increase the tracking performance, we estimate the friction disturbance torque and then compensate it using MATLAB/SIMULINK. We addressed the frictional disturbance torque in both the channels of the gimbal system using observer-based controller design and stabilize the gimbal-based robotic eye. To show the Saccade like the human eye, we performed simulations for tracking control. The designed controller is robust against frictional disturbance. The simulation results illustrate the stabilization and tracking control of the gimbal-based robotic eye and the estimation and compensation of frictional disturbance. This will be a valuable step towards humanoid robots.