We present a control framework for robot systems to perform dynamic manipulation tasks, with robotic batting as an example. A properly preplanned trajectory for manipulators is recognized to be important to increase the probability of success. A nominal path for a manipulator to travel in a swing motion is computed while optimizing the stochastic dynamic manipulability measure to cope with the unpredictable deviation of the ball. A nominal trajectory along the path is then planned in a minimum-time fashion for the benefit of catching the ball as accurately as possible, the behavior of which is observed visually and estimated on-line. A control structure consisting interpolatively of a path velocity controller to follow the nominal trajectory and a resolved-acceleration-natured ball-deviation compensating controller, together with a ball motion prediction, is proposed. The effectiveness of the proposed approach is examined using computer simulations. The strategy provided in this paper will be useful to ma...