The invention provides a method for establishing a data mining automatic feedback system, used for solving the problem that manual feedback is continuously needed to implement adjustment and optimization during an existing data mining implementation process. The data mining automatic feedback system established according to the method comprises a data segmentation module, a result assessment module and a parameter adjustment module, wherein the three modules work cooperatively to form feedback, and automatically adjust and optimize a parameter of a data mining algorithm, so that more labor cost is saved than traditional data mining. The algorithm parameter is effectively, accurately and automatically adjusted by an automatic parameter adjustment algorithm, parameter adjusting efficiency is effectively improved, and actual waste caused by excessively traversing a parameter value range during an automatic process is reduced.