The intent of this research work is to emphasize upon online social network mining referred as link prediction applied on undirected online social network represented through undirected graphs. The main aim of link prediction is to identify future missing links between nodes of the network based on previous relations that is network topology. The existing state of the art link prediction techniques also have been studied as to understand their limitations. It has been discovered that link prediction helps in addressing the issues and challenges of various domains such as recommendation of future friends. Although there are issues related to precision and scalability. In case of precision the diversity of online social networks in their nature is a challenge, because link prediction techniques used for forecasting future missing links in a specific network will not going to precisely predict the links between the nodes in another type of network. The properties of friendship network are different than co-authorship network.