The Rock Mass Rating (RMR) system is a classification based on the six parameters which was defined by Bieniawski. This system may possess some fuzziness in its practical applications. For example, experts mostly relate discontinuities and ground water conditions in linguistic terms with approximation. Descriptive terms vary from one expert to the other, while in the RMR system; values which are related to these terms are probably the same. The other hand, sharp transitions between two classes create uncertainties. So it is proposed to determine weighting intervals for discontinuities and water condition. Two fuzzy models based on the Mamdani algorithm were introduced to evaluate proposed weights, so that the first fuzzy model includes 55 scores using fuzzy model and the remained scores which are related to discontinuities and ground water conditions are obtainable by the RMR system. But the second fuzzy model obtains all scores of the RMR system using fuzzy model. Results of fuzzy models are adapted with actual RMR, but second fuzzy model predicts more acceptable results, because it has the ability to use qualitative terms in fuzzy state. But first fuzzy model uses descriptive terms in classic state. So it seems, proposed weighting intervals can manage fuzzification of discontinuities and water conditions.