The number of known complexes recalled by the existing algorithms of clustering in proteinprotein interaction network is very limited. To solve this problem, a new distance measurebased algorithm for identification of protein complexes, named IPCDM, is proposed based on our discovery that most of the shortest paths between proteins complexes are no more than two. A new seedextension model is also proposed to improve the precision of protein complexes discovery. Experiment results on yeast protein interaction network show that more known protein complexes are recalled by IPCDM than by other typical algorithms: MCODE, ENSC, CFinder, LCMA and DPClus.