This chapter covers link prediction through its principles, methods, and applications. It forecasts potential future connections and identifies currently unknown links across both temporal and spatial dimensions. Link prediction has become a prominent research area, expanding its techniques by integrating various models. This chapter presents three key models, illustrating their interconnections. Furthermore, advancements in neural networks and deep learning have led to the creation of graph-based models that combine network structures and topologies. Link prediction is widely applied in fields like social network recommendations (e.g., Weibo, QQ, and Twitter) and in predicting node types in known networks, such as detecting spam emails or forecasting criminal behavior. Despite its broad applications, link prediction remains an active research topic in social networks.