Abstract: Based on Fisher discriminant theory, the Fisher Discriminant Analysis (FDA) model was established for classifying mine-lane surrounding rocks. Six factors such as roadway depth, roadway span, coefficient of mining influence, strength of surrounding rock, thickness of loose zone, situation of joint development were selected as the indicator system of the FDA mode. Discriminant function for classifying mine-lane surrounding rocks is obtained in accordance with the measured data as training samples. One weak indicator, which is thickness of loose zone, was removed from the indicator system through analysis, and then we got four discriminate functions. In order to verify the validity of the FDA model, 15 groups of measured data are used as samples to make training, the backward substitution method was adopted to verify the effectiveness of the model, with the misjudgment rate of 0, then, the FDA model was applied to engineering data in the same area, and the effect is good. The results show that the FDA model is simple and accurate; it is an effective method for the rapid identification of the classification of mine-lane surrounding rocks.