Reduction of fractures in the minimally invasive (MI) manner can avoid risks associated with open fracture surgery. The MI approach requires specialised tools called percutaneous fragment manipulation devices (PFMD) to enable surgeons to safely grasp and manipulate fragments. PFMDs developed for long-bone manipulation are not suitable for intra-articular fractures where small bone fragments are involved. With this study we offer a solution to potentially move the current fracture management practice closer to the use of a MI approach. We investigate the design and testing of a new PFMD design for manual as well as robot-assisted manipulation of small bone fragments. This new PFMD design is simulated using FEA in three loading scenarios (force/torque: 0N/2.6Nm, 75.7N/3.5N, 147N/6.8Nm) assessing structural properties, breaking points, and maximum bending deformations. The PFMD is tested in a laboratory setting on Sawbones models (0N/2.6Nm), and on ex-vivo swine samples (F=80N±8N, F=150±15N). A commercial optical tracking system was used for measuring PFMD deformations under external loading and the results verified with an electromagnetic tracking system. The average error difference between was 0.5mm, within the accuracy limits of the respective tracking systems. Final results from reduction manoeuvres performed both manually and with the robot assistance are obtained from 7 human cadavers with reduction forces in the range of (F=80N±8N, F=150±15N, respectively). The results show that structurally, the system performs as predicted by the simulation results. The PFMD did not break during ex-vivo and cadaveric trials. Simulation, laboratory, and cadaveric tests produced similar results regarding the PFMD bending. Specifically, for forces applied perpendicularly to the axis of the PFMD of 80±8N deformations of 2.8mm, 2.97mm, and 3.06mm are measured on the PFMD, while forces of 150±15N produced deformations of 5.8mm, 4.44mm, and 5.19mm. This study has demonstrated that the proposed PFMD undergoes predictable deformations under typical bone manipulation loads. Testing of the device on human cadavers proved that these deformations do not affect the anatomic reduction quality. The PFMD is, therefore, suitable to reliably achieve and maintain fracture reductions, and to, consequently, allow external fracture fixation.