The study investigated the effect of exhaustive physical exercises upon the performance of a manual precision control task. It was hypothesized that the more related the limb-segments were to that used in the task, the more significant would the decrement be in performance. Four types of exercises were used in this experiment: the whole-body, the fingers, hand-finger and the arm. The motor task was the steadiness test in which the hand and fingers were required for precision control. Sixteen right-handed male subjects were used. Results showed that local exercise of the fingers or the hand produced significantly more decrement in term of longer time than that from the task exercise of either the whale-body or the arm. No difference in task time was observed between exercises of the hand and the fingers or between those of the arm and the whole-body. No difference was found for the error measurement of the control task. No learning effect was found in any of the conditions with or without exercises.