The mining and analysis of community in complex networks is an important issue in many domains and disciplines. In this paper, focus on the sparsy Communication Matrix of the massive data, we suggest a fast mining algorithms based on a heuristic function to merge called KFN algorithm, and also show out the MapReduce model of this algorithm. With experiment focued on the DBLP paper co-network data sets, we conclude that distributed mining model and the merging strategy based on the heuristic function can improve the time efficiency on community mining.