Abstract: Aiming at the subjective arbitrariness of group decision making and linguistic fuzzy uncertainties in the risk assessment of dam breakage of tailings pond， a multi-level fuzzy comprehensive assessment model of dam breakage risks of tailings pond based on the extension analytic hierarchy process （ EAHP） is established. The extension interval judgment matrices have been used to objectively and comprehensively describe the vague opinions and preferences of experts on the evaluation index weight of the tailings dam breach risk. The deviation between the opinions of each expert and the expert group has been used to calculate the importance coefficient of each expert. Then the index weight coefficients [ωj] have been determined after the consistency coefficient [α] and [θ] have been introduced to deal with extension interval judgment matrix R.Doing by this， the effects of all the experts in the group decision- making of the dam break risk about the tailings pond have been highlighted， the consistency test process of extension interval judgment matrices have been simplified and the reliability of index weight has been improved. The multi-level fuzzy comprehensive evaluation model of the dam breaking risk about the tailings pond has been applied to analyze the dam break risk of the Yanjiawan tailings pond. The evaluation results show that the risk level of the tailings pond is safe， and the dam body instability and flood overtopping have a great impact on the safety of the tailings pond， which is consistent with the actual the daily management situation of Yangjiawan tailings pond. This method can be used to effectively solve the problem of weight group decision making in the risk analysis of tailings dam break， avoid the loss of expert opinions， reduce the adverse impact of the randomness of expert's subjective judgment on the evaluation results and improve the effect of the risk analysis of tailings dam break.

Key words: tailings pond, risk of dam break, EAHP, degree of membership, fuzzy comprehensive evaluation