Paralympic sports are growing more popular. Besides the dedicated training of the athlete, technology is crucial to empower amputees to perform at their highest level. The recent carbon fibre Running-Specific- Prosthesis (RSP) have energy stored and return capabilities that allow runners with amputations to perform almost as able-bodied. However, due to the difference in power output between a biological ankle and a RSP, unilateral transtibial (UTT) amputees need to adjust their biomechanics to an asymmetric pattern. It is known that the stiffness of the blade has a great influence on the performance of the athlete. The main challenge was to find a method to prescribe the optimal stiffness for a particular athlete, that allows him/her to performthe best in a race. The current approach to advice a RSP is based on the bodyweight of the athlete and the coaches and athletes wishes. Nevertheless, as a result of the insufficient number of subjects and the difficulty to performa randomised control trial, there is limited evidence about the UTT athlete capability of adapting their muscle activity to different prosthesis stiffness and the optimal RSP stiffness for them. In order to investigate this, two main goals for this master thesis project were raised: to implement a musculoskeletal model of an UTT amputee athlete wearing a RSP and predict its optimal running biomechanics through predictive forward dynamic simulations; to find the optimal athlete-RSP stiffness combination that maximizes the running performance. Five different stiffness of a Flex-Run ¨Ossur (Reykjavík, Island) RSP were modelled and simulated for the maximumvelocity the model could reach. Results: firstly, the model could perform better with a middle-class RSP category. It enhanced the hip muscles to exert more power in the blade during the first part of the stance phase. However, the prosthetic leg generated 41.2% lower total average power, including the RSP power, than the intact leg. The RSP made up 24.8% of the total prosthetic leg power. Secondly, the intact leg benefited from an improved push-off of the prosthetic leg having a favourable landing that allowed to exert more power and propel the body into longer flight time than with low-class RSP categories. Still, the top speed of the model was far from what athlete can achieve, but the motion and kinetic data were comparable for low running speeds. Therefore, it could be concluded that a reasonable prediction of the optimal RSP stiffness for running at about 4.6 m/s was achieved. The presented biomechanical model could potentially be used to assist coaches and athletes to have a better idea of themost suitable RSP stiffness.