The invention discloses a selection method for a collaborative feedback set and an eNodeB. The eNodeB receives reference signal quality parameters of each cell reported by user equipment (UE); corresponding judgment thresholds of each non-serving cell are determined respectively according to the ratio between reference signal quality parameters of each cell and/or serving cell business volume and the business volume of non-serving cells; the eNodeB respectively calculates the differences between the reference signal quality parameters of the serving cells and the reference signal quality parameters of the non-serving cells, the differences obtained by calculation are respectively compared with the corresponding judgment thresholds of each non-serving cell, and if the differences obtained by calculation are smaller than the judgment thresholds, the corresponding non-serving cells are placed in the collaborative feedback set. The judgment thresholds are dynamically adjusted according to the ratio between the reference signal quality parameters and/or serving cell business volume and the business volume of the non-serving cells, compromise between feedback expense and collaborative transmission gain is achieved, and the large collaborative transmission gain is obtained with the small feedback expense.