The variation in each patient's hand appearance and function, such as muscle atrophy, stiffness, and tension, brings difficulty in the molding of the auxiliary splint used to correct and aid in the hand function recovery of clinical stroke patients. The hand rehabilitation assistive devices for stroke patients currently sold in the market are effective in the rehabilitation of the patient's handgrip. However, there are still a lot of ground for improvement for clinical practice-the limited customization, the high unit price, and the difficulty in wearing. This research work successfully developed a man-machine interface system for clinical use mainly utilizing the commercially available 3D printers. The system is operated by scanning the limbs not affected by the stroke with the use of a portable 3D scanning platform. With the scanned 3D image, a design for the splint of the clinical patient is produced through the surface reconstruction technology, 3D image mirroring, and hollow designs. The completion of this system provides great help in the manual production process of the assistive devices and the customization of splints for most clinical stroke patients.