The present research uses community discovery algorithm to identify and extract the community structure inside the giant branches of domain knowledge networks. Through the calculation and analysis on inscape of knowledge communities, export of knowledge communities and lineage between communities, the time series dynamic tracking and analysis on evolution pattern and knowledge lineage characteristics of knowledge communities are carried out. The research results show that the domain knowledge networks are more non-connected in the process of evolution; the more mature the domain knowledge develops, the more frequently the knowledge cross-fusion between communities; the cross-fusion knowledge lineage mostly occurs between the knowledge communities within the giant branch.