Emotion is defined as a physiological and psychological state that encompasses human thoughts, behaviors, and feelings. This phenomenon is also regarded as a spontaneous physiological and psychological response generated by the human body to external stimuli. Given the established correlation between electroencephalogram signal and cerebral activity, it is possible to extrapolate the emotional state of subjects by means of electroencephalogram signal analysis. In this review, emotion models, datasets, and popular machine learning and deep learning methods in recent years used in emotion recognition research were summarized. In addition, the research progress of emotion recognition based on electroencephalogram signal was reviewed, with the aim of assisting subsequent researchers in understanding developments in electroencephalogram signal domain and offering insights for addressing clinical challenges in emotion recognition.