Under the hesitant fuzzy environment, considering the decision makers' psychological behavior, a method based on regret theory and evidence theory was proposed to cope with Multi-Attribute Group Decision Making (MAGDM) problems that the attribute value is hesitant fuzzy information, the attribute weights and probability information of situation are completely unknown. First, evidence theory was utilized to calculate the probability information of the states. Then, based on the interval fuzzy matrices, the estimation of t-distribution and the score function matrices, the utility values of attribute values were determined. Moreover, by using regret theory, the decision makers' perceived utility values were obtained. The overall perceived utility of each alternative was acquired on the basis of the weighted arithmetic mean. After that, all the alternatives were ordered. Finally, the proposed approach was applied to a numerical example about the selection of enterprise. The experimental results show that the proposed method can get the same results as the existing methods, but only a small number of parameters needed to be considersed in the decision process. The results of comparative analysis domonstrate that the decision making results obtained by the proposed method are reasonable and reliable, which can reflect the actual situation.