In order to solve the large group emergency decision making problem with the characteristic of multi-department and multi-index, a new decision making method is proposed. Firstly, the similarity formula of intuitionistic fuzzy numbers is constructed, which is taken as the clustering formula for expert preference clustering operation. Then, aiming at maximization of the distinction degree over index data for each department, and index weights for each department are obtained. Based on the degree of differentiation over department information, combining with maximum entropy principle, department weights are calculated, and then all alternatives are ranked. Finally, an example is used to illustrate the feasibility and effectiveness of the proposed method.