Robot and functional electrical stimulation (FES) techniques have been used for improving motor functions after stroke respectively. In this work, we aimed to examine the combined effects from Robot and FES in post-stroke wrist rehabilitation. We developed a new electromyography (EMG)-driven FES-robot system with a control algorithm, which was interactive to the voluntary motor inputs from persons after stroke. Two hemiplegic subjects with chronic stroke were recruited to test the EMG-driven FES-robot system. The results demonstrated that the integrated system was effective in improving the tracking performance represented by root mean squared error (RMSE) and root mean squared jerk (RMSJ) of tracking trajectories in programmed wrist tracking tasks assisted with different proportions of the interactive supports from the FES and robot parts respectively. This combined system may have the potential to be applied in stroke rehabilitation.