An active end effector based force control system for robotic deburring is successfully implemented using a PUMA-560 robot. The system goal of a controlled chamfer depth with minimum surface roughness is achieved by minimizing the normal chamfering force variance online. Several force control algorithms are evaluated based on this objective. The control laws are designed based on models combining a deterministic plant with a stochastic disturbance which are identified from experimental data. Simulation results are verified by real-time force control experiments. Performance comparisons are made based on the force variance and surface roughness achieved by each controller. The 6 step extended horizon controller is shown to achieve the best overall performance.