The Human hand is versatile in its interaction. Imitating the versatilities and human dexterity in robotic design is a huge challenge, and requires a great depth of understanding of the human upper limb physiology and its robotic equivalent. A huge amount of research worldwide is being carried out in order to develop a human hand like prosthetic which can provide natural haptic functionality. This paper provides an insight to the development of a bionic hand which performs hand opposition and reposition actions (clasp and release) based on real EMG signals from a below elbow amputee. It also provides an understanding of design involved in development of the robotic model of the hand, the electronics behind it and finally the signal processing technique behind classification of EMG signals to make the bionic hand perform the desired haptic function. The paper indicates that Artificial Neural Network (ANN) and Random Forest method is the best technique that can be used to classify actions indicated through real time EMG signals.