Detecting scientific research community based on CNKI’s academic papers could provide personalized recommendation service for reading documents of the majority of scholars. This paper put forward a scientific research community detection algorithm, firstly, established the scientific research paper coauthor network by Pajek, regarded Diningtable partners and Sampson of network common data set as testing data set to have a performance comparison and analysis on the scientific research community detection algorithm and classical GN algorithm, verified that the performance of the scientific research community detection algorithm was more superior and finally found scientific research community structure using the algorithm, and the experimental results showed that the effect of community division was pretty good.