Abstract: Unascertained measurement theory was applied in tunnel surrounding rock classification, unascertained average clustering method was raised and unascertained average clustering model of tunnel surrounding rock classification was set up.Unit weight, tensile strength, sound degree coefficient of rock was selected as the evaluating parameters of classification model.Using 14 kinds of measured data of rock to build the unascertained measurement function of classification of judging indexes, average value of each classification sample index and single index measure matrix were obtained; weight of each index confirmed according to the information entropy theory, determine classification of rock was done by using credible degree recognition criterion.Classification forecasting was done on tunnel surrounding rock of mining area by using this model.Prediction results are closely with measured results after comparison, and it offers a new way for tunnel surrounding rock classification.

Key words: Tunnel surrounding rock, Classification, Unascertained mean rating method