Socio-cognitive skills are crucial for successful interpersonal interactions. Two particularly important socio-cognitive processes are emotion perception (EP) and theory of mind (ToM), but agreement is lacking on terminology and conceptual links between these constructs. Here we seek to clarify the relationship between the two at multiple levels, from concept to neuroanatomy. EP is often regarded as a low-level perceptual process necessary to decode affective cues, while ToM is usually seen as a higher-level cognitive process involving mental state deduction. In information processing models, EP tends to precede ToM. At the neuroanatomical level, lesion study data suggest that EP and ToM are both right‐hemisphere based, but there is also evidence that ToM requires temporal-cingulate networks, whereas EP requires partially separable regions linked to distinct emotions. Common regions identified in fMRI studies of EP and ToM have included medial prefrontal cortex and temporal lobe areas, but differences emerge depending on the perceptual, cognitive and emotional demands of the EP and ToM tasks. For the future, clarity of definition of EP and ToM will be paramount to produce distinct task manipulations and inform models of socio-cognitive processing.