Underactuated grippers aim to simplify the control strategies for performing stable grasps due to their inherent shape adaptability. While at the beginning, the main research area was focused on developing human-like robotic hands for disabled people, in the last years, a new eld of application appeared with the constant evolution of the industry: the implementation of a single underactuated gripper as a replacement of diverse dedicated fully-actuated grippers. However, two main issues are restraining its use: the stability of the grasp and the speed of performance. The rst is an active topic as all underactuated grippers need to ensure the stability of the grasped object through an adequate kinematic design, while, the latter is not widely treated as there weren't many application elds where high-speed was required and, at the end, the quasi-static analysis must be also ensured. For this reason, the present research work has been focused on the speed of the grasping. In the rst place, an introduction to underactuated hands is made, and is followed by two main stability criteria. Then, the development of a model for an underactuated nger that allows analyzing the complete grasping sequence at high-speed along with a collision model are presented. Following, a design-based analysis to simplify the model is performed, and the graspstate volume tool is introduced in order to inspect the impact of the design variables on the proposed criteria. In the last chapter, an optimization over the design space is performed and a design is chosen, crosschecked with ADAMS software and prototyped. Finally, an overview remarking the strengths and gaps in the research is presented in the form of conclusions, and closing them, future works that could be interesting to develop.