Hard computing based optimization algorithms usually require a lot of computational resources and generally do not have the ability to arrive at the global optimum solution. Soft computing algorithms on the other hand negate these deficiencies, by allowing for reduced computational loads and the ability to find global optimal solutions, even for complex cost surfaces. This paper presents fusion of soft computing or control technique of genetic algorithm (GA) and hard computing technique of proportional integral derivative (PID) control with application to prosthetic hand. An adaptive neuro-fuzzy inference system (ANFIS) is used for inverse kinematics of the three-link index finger, and feedback linearization is used for the dynamics of the hand and the GA is used to find the optimal parameters of the PID controller. Simulation results with practical data shows good results for the prosthetic hand to hold a square object with a two-link thumb and index finger.