Abstract：The occurrence and intensities of rockburst are influenced by various factors and the influence degrees of these factors vary with different projects. To evaluate the various factors on rockburst，the concepts，definitions and methods of variable fuzzy sets are established based on fuzzy sets theory. Then，the formulas of difference functions and relative membership degree，model of BP neural network with fuzzy optimization，equation of rank feature values are presented，and relevant equations are analyzed. On basis of indexes standard intervals of rockburst with a series of underground rock projects in home and abroad by the presented methods，occurrence and intensities of rockburst with rank feature values are predicted. The predicted results agree well with the practical records，which shows that the proposed method is effective.
