This paper presents preliminary results from a virtual reality (VR)-based system for hand rehabilitation that uses a CyberGlove and a Rutgers Master II-ND haptic glove. This computerized system trains finger range of motion, finger flexion speed, independence of finger motion, and finger strength using specific VR simulation exercises. A remote Web-based monitoring station was developed to allow telerehabilitation interventions. The remote therapist observes simplified versions of the patient exercises that are updated in real time. Patient data is stored transparently in an Oracle database, which is also Web accessible through a portal GUI. Thus the remote therapist or attending physician can graph exercise outcomes and thus evaluate patient outcomes at a distance. Data from the VR simulations is complemented by clinical measurements of hand function and strength. Eight chronic post-stroke subjects participated in a pilot study of the above system. In keeping with variability in both their lesion size and site and in their initial upper extremity function, each subject showed improvement on a unique combination of movement parameters in VR training. Importantly, these improvements transferred to gains on clinical tests, as well as to significant reductions in task-completion times for the prehension of real objects. These results are indicative of the potential feasibility of this exercise system for rehabilitation in patients with hand dysfunction resulting from neurological impairment.