This chapter focuses on computational theories of emotion. Emotion is much more difficult for researchers to study in laboratory settings due to its subjective nature. Despite its long history, compared to the studies of vision, audition, memory, and many other mental processes, emotion research still faces several major challenges. These challenges include: the definition of emotion is imprecise and many related terms are used interchangeably; there is profound inconsistency between subjective reports of emotion and objective measures such as functional magnetic resonance imaging (fMRI) or electroencephalogram (EEG); and existing analyses are largely correlational and lack mechanistic or computational approaches. Over the past few decades, there have been considerable developments in the field of computational modeling of emotion. Most of these developments come from artificial intelligence, connectionism, Bayesian modeling, and neuroeconomics. The ideas are deeply rooted in the cognitivist tradition of emotion research as well as the embodied view of emotion.