There have been many researches on the joint limit algorithm exploiting kinematic redundancy in kinematically redundant manipulators. Existing joint limit algorithms tend to minimize the distance between the current joint angle and the middle angle of the given joint range. However, strictly speaking, these algorithms control the joint angle to come close to the middle angle, but sometimes this algorithm does not fully utilize the entire motion range of joints and consumes relatively large energy. Therefore, this paper proposes a new joint limit algorithm that not only prevents the joint limit violation, but also utilizes the wide motion range of joints. Two exemplary manipulator models having kinematic redundancy are employed to show the feasibility of the proposed algorithm.