Different combinations of segment height, route spacing and breaking interval have a significant effect on the indicators of the depleted ore loss, thus affecting the mining economic benefits. In this paper, the uncertainty measurement theory and the AHP(Analytical Hierarchy Process) are used to establish an evaluation model of the unit concentrate earnings. The evaluation model takes account of various factors that affect the unit concentrate earnings, such as the dilution rate, the recovery rate, the grade of the rock mixed with waste rock and the grade of tailings. The measured data are used to create a single index unascertained measurement matrix, then calculate the weight of each index by the AHP and evaluate the unit concentrate earnings according to the credible degree recognition criteria, thus determine the preferred caving step. In examples, the unit concentrate earnings are optimal when the breaking interval is 4.8m for the 440 stratified of Dahongshan Iron, and the step for the optimal breaking interval is determined.