This paper considers the design and implementation of a visual servo system for catching a flying ball using an active stereo vision system, a static vision system, and an omni-directional wheeled mobile robot. These multi-camera vision systems are used to track the flying ball and guide the omni-directional mobile wheeled robot to catch it. The dynamic model of a flying ball and Kalman filter are used to mitigate measurement noise, estimate the position and velocity of the flying ball, and predict its future trajectory. The mathematical model of the omni-directional mobile robot is derived to facilitate the control design. Trajectory tracking control of the mobile robot is done by combining feedback linearization and proportional-integral-derivative (PID) control. Finally, the experimental setup is constructed, and control laws and image processing algorithms are implemented through digital signal processors. The effectiveness of the designed system is validated through experimental studies. The results show that the developed vision systems are able to track a flying ball and guide and navigate the robot to catch it.