Social network analysis has been widely used in studies of animal behavior. Application of social network analysis to animal behavior can identify the key individual in a social group, uncover the formation mechanisms of social associations, improves our understanding of the origin of sociality, and identify the links between individual behavior and population-level dynamic patterns by quantifying specific attributes (e.g. centrality, centralization). Here, we first present a brief history of social network analysis, and a description of basic concepts. Then, we discuss different approaches for inferring social networks from observing data, and introduce some social attributes. Thirdly, we highlight the necessity and approach of constructing null models. Finally, we review previous applications to animal behavior, propose that future studies should focus on both the dynamics of physical contact (e.g. fighting, grooming) and nonphysical contact (e.g. vocal communication), and on mixed-species societies, so that it will further our understanding of the ecological function of social behavior and cooperation theory.