WeChat group as a new social relationship organization model, is widely used in work, study and life. People exchange ideas and transmit knowledge through different WeChat groups. To study the differences of knowledge sharing within WeChat Group in different situations can solve the problem of inefficient knowledge sharing in WeChat Group and provide support for diversified knowledge governance. This paper researched the effects of structural capital, relational capital and cognitive capital on the explicit and tacit knowledge sharing within the work, learning and living WeChat group based on the theory of social capital by random graph model(ERGMs). The results show that the key factors affecting tacit knowledge sharing in the work scene are individual centricity, interaction intensity, identity, reciprocity and trust. The key factors that affect explicit knowledge sharing in the work scene are individual centrality, identity, reciprocity, trust, and common language. The key factors affecting tacit knowledge sharing in learning scenes are individual centrality, identity, reciprocity and trust. The key factors that affect explicit knowledge sharing in learning scenarios are individual centrality, interaction intensity, reciprocity, trust and common language. The key factors that affect the sharing of tacit knowledge in life scenes are the intensity of interaction, identity, trust and common language. Trust and common language are the key factors that affect explicit knowledge sharing in life situations.