This paper surveys the theories of social preferences. Social preferences are based on that people not only care about their own well-being, but they have a certain concern with payoffs and/or actions of others. We classify two approaches: distributional and intention-based models, and later discuss models that combine both theories. In order to provide a better illustration of the discussed models, we derive predictions of these models for two classic experimental protocols: ultimatum game and public good game with punishment. These predictions are compared with the stylized facts of these two games.