<p>Given the increasing prevalence of post-stroke hand dysfunction and the limitations of existing rehabilitation robots–rigid structures offering high precision but poor compliance, and flexible systems providing comfort but low control accuracy, this study proposes a novel rigid-flexible coupled hand rehabilitation robot based on a “unit selection” design strategy. By analyzing the kinematic characteristics of the human hand, a finger equivalent model was established using the D-H parameter method. The design integrates rigid constraint tracks with a tendon-driven system and micro-motors to achieve synergistic optimization of precision and compliance. Dynamic simulations using ADAMS software validated the smooth motion and safe joint angle variations of the mechanism. Prototype experiments with 3D-printed components and an image-capture system confirmed the feasibility of the design and the accuracy of rehabilitation trajectories. Results demonstrate that the proposed robot effectively combines the control advantages of rigid structures with the compliance of flexible systems, offering a new solution for enhancing hand rehabilitation outcomes.</p>