Abstract: Two-tier heterogeneous networks composed of femtocell and macrocell can improve the system capacity effectively. A distributed IA algorithm is proposed based on clustering. The proposed algorithm achieves clustering of femtocell network, which based on the mechanism of the edge weights of graph theory and the criterion of maximum the edge weights. Moreover, the macrocell user, which causes the strongest interference, is selected to form a whole with a certain cluster. Finally, the DIA technology is used to eliminate the interference of the whole. The proposed algorithm guarantees the performance of femtocell network system, meets the requirements of the feasibility conditions of dense deployment case effectively. Simulation results show that the proposed algorithm has better performance of channel capacity and is suitable for the actual two-tier heterogeneous networks.