Facial expression is a basic form of emotional communication. Correctly recognizing other’ facial expression plays crucial role in social interaction. However, facial expressions usually do not appear in isolation but are often embedded in some emotion-loaded situations. That means, efficient emotion perception from a face could not rely on facial expression alone, but could be affected by the contexts. The purpose of this article was to review the studies addressing the contextual influence on recognition of emotional facial expressions. There are two different types of promotion effect of emotional context on facial expression recognition: (1) emotion-congruent effect in visual modal. When a facial expression appeared in visual context, the emotional context will facilitate the recognition of facial expression with the congruent emotional valence. This effect was demonstrated by decreased response time and increased accuracy at behavioral level, and increased amplitude of N170 and increased activation in left fusiform gyrus at neural level; (2) integration effect of cross-modal emotional information. Studies found that, context from other modals such as auditory and olfactory context could also facilitate the recognition of facial expressions. The inhibition effect of emotional context on facial expression recognition consists of two aspects: (1) emotional conflict effect, which appears when the emotional valence of the target face was incongruent with the distractor context. Studies concerning emotional conflict mainly adopted word-face Stroop and emotional Flanker paradigms. Moreover, the emotional conflict effect could also be induced by cross-modal emotional distractor (e.g., laughing at a fearful face). Neural studies showed that the emotional conflict effect was associated with increased amplitude of the N400 component and a neural network (i.e., amygdala, dorsomedial/dorsolateral prefrontal gyrus and dorsal anterior cingulate for conflict monitoring and rostral anterior cingulate for conflict resolution); (2) semantic interference effect. This effect was mainly explored by semantic satiation paradigm, i.e., when a word representing a certain emotional category (e.g., “anger”) was satiated and become semantically inaccessible, the processing to the associated emotional facial expressions (e.g., angry faces) would be interfered. The mechanisms underlying influential effect of emotional context on neutral and ambiguous faces include: (1) inductive effect of emotional context. Studies showed that, emotional context could bias the neutral faces to certain affective tone congruent with the context. While for ambiguous faces, the emotional context could disambiguate them to show processing features similar with the emotion of the context; (2) subliminal affective priming effect. Studies showed that briefly presented emotional context could also affect the recognition of neutral or ambiguous facial expressions. Finally, the article summarized the different mechanisms in the existing literature: (1) the promotion effect of emotional context on emotional facial expressions, and the inductive effect of emotional context on neutral and ambiguous facial expressions could be generalized by emotion-congruent effect; (2) the inhibition effect of emotional context on emotional facial expression recognition includes two dimensions: emotional dimension (i.e., emotional conflict effect) and conceptual dimension (i.e., conceptual interference effect). Furthermore, some suggestions for future studies were proposed. Firstly, future studies should investigate the influence of multi-modal emotional context on facial expression recognition. Secondly, future studies should extend to explore the social emotions (e.g., puzzled) and self-conscious emotions (e.g., self-abased). Thirdly, future studies should explore the preference judgment of facial expressions in emotional context, other than the categorization of emotion.