To demonstrate the hardware realizability and efficacy of the quadratic program (QP) based methods for solving the nonrepetitive problem, this paper proposes a novel repetitive motion planning and control (RMPC) scheme and realizes this scheme on a physical planar six degrees-of-freedom (DOF) push-rod-joint (PRJ) manipulator. To control the PRJ manipulator, this scheme considers variable joint-velocity limits and joint-limit margins. In addition, to decrease the errors, this scheme considers the position-error feedback. Then, the scheme is reformulated as a QP problem. Due to control of the digital computer, a discrete-time QP solver is presented to solve the QP problem. For comparison, both of the nonrepetitive and repetitive motions are performed on the manipulator to track square and B-shaped paths. Experimental results validate the physical realizability and effectiveness of the RMPC scheme.