The paper presents a systematic methodology for designing tracking control of robotic motion. The methodology is completely based on numerical methods and corresponding recurrence relations. It produces recursive control laws which force the system to exhibit the desired error dynamics. A few explicit, easily realizable control algorithms are given and compared with the computed-torque scheme. Closed-loop stability and bounds on system uncertainties are investigated in detail. Control algorithms that use only position measurements are also analysed. A benchmark problem is solved to illustrate the use of proposed algorithms.