Abstract: In consideration of the imperfection, randomness and ambiguity of the information for stability analysis of highway slopes, an improved fuzzy evaluation method is put forward to assess the stability of highway slopes based on the maximum entropy principle and the engineering fuzzy set theory. In this method, the comprehensive empowering method is adopted to determine the weight of index. The procedure makes full use of the information of the factors influencing stability of highway slopes, and uses the generalized weighted distance to represent the difference between slope sample and standard sample. The relative membership degree is obtained by conditions’ optimization, the grades of the slope samples are identified according to the weighted average principle. The procedure is applied to Changde-Jishou and Jishou-Huaihua highway in Hunan province to analyze the stability of rock slopes. It is found that the assessment results agree well with the results obtained with the fuzzy comprehensive evaluation method，the attribute recognition method and the practical condition of slopes, showing that the new procedure is more reasonable. Furthermore, the Shannon entropy value of the model is smaller, which indicates less uncertainty and higher reliability of the evaluated results.

Key words: rock slope, maximum entropy principle, relative membership degree, comprehensive empowering, weighted average principle, stability evaluation