A classification and aggregation problem on large-scale group decision making is studied based on the hesitant fuzzy linguistic terms. Specifically, a method to measure the similarity of two hesitant linguistic sets is proposed considering both hesitancy and distance. Then, the preliminary expert classes are generated using the netting method based on a similarity matrix, and an acceptable level is set to help make a second classification. Through a classification accuracy index, the final classes are obtained. Furthermore, two aggregating frames are constructed respectively for the information within one class and between the classes. The proportional linguistic groups are obtained through combing the information within one class. On that basis, the degree of reliability of each class is calculated and the class weights are determined based on the class reliability and the percentage of expert number in one class to the total number to aggregate the information among classes. Additionally, the expected values of the alternatives are calculated to make a selection. Finally, a case is given to illustrate the effectiveness and feasibility of the proposed method.