Most community discovery methods determine the best community according to network topology and edge density, however, they have very high computational complexity, and are very sensitive to the form and type of the network. To solve these problems, we propose an interactive optimization algorithm based on dynamic nodes adaptive incremental model for micro-blog communities. By optimizing the interaction among members in each community, the algorithm efficiently searches the candidates of the optimal community without traversing all the nodes by using the greedy algorithm. The model allows rapid and accurate measurement of interaction difference within community and across communities. Finally, simulations on the data grabbed from benchmark networks and Sohu microblog platform show that the proposed algorithm outperforms other algorithms in recall rate, accuracy, computation time, and network coverage rate.