The field of autism research is dominated by the mind-blindness theory, which is based on the mentalistic supposition and the supposition of universality. These assumptions advocate that other minds are hidden away, closed in, behind the overt behavior that we can see, and we understand others mainly by this mind-reading way. Interaction Theory refuses these two suppositions, and proposes that intersubjectivity involves three levels: primary intersubjectivity, secondary intersubjectivity and narrative competency based on the evidence of development psychology and phenomenological research. On the level of ontogeny, these three have been in succession, and in the social interaction level of mature individuals they are mutually beneficial, but the foremost way of understanding others is the intuitive social interaction which is on the basis of sensory-motor. There are obstacles of autistic children with the primary intersubjectivity and the secondary intersubjectivity. Actually, Both scientific evidence from developmental psychology and phenomenology suggest that theory of mind itself is not a good explanation of non-autistic intersubjective experience. In the view of Gallagher, If theory of mind can not offer an acceptable account of our normal interaction with others, then the lack of a theory of mind does not offer a good account of the problems of autistic children. In order to explain the problem of social interaction of autism, Gallagher outlined the interaction theory, which incorporates evidence from both developmental and phenomenological studies to show that humans are endowed with important capacities for intersubjective understanding from birth or early infancy. And in adulthood these abilities don’t disappear, but continue to play the role, and only in a special situation which the ability is not enough to explain the more advanced narrative ability will play a role. From perspectives of neuroscience and phenomenology, interaction theory consider the effects that both sensory-motor problems and problems of central coherence may have on primary and secondary intersubjectivity as well as their connections to the social symptoms. Certainly, the interaction theory is facing many challenges now, and we also look forward to provide a unified research program which integrate developmental, phenomenological, enactive, and dynamical approaches of social interaction together.