Abstract: To solve the credit risk evaluation problem in which the attribute weights are completely unknown and the data are multidimensional time series,a method of credit risk evaluation modeling based on the multi-attribute decision making and fuzzy clustering is presented.This method uses deviation-based maximization method and quadratic programming model to determine comprehensive weights of index attribute.The evaluated samples on each time point are scored by multi-attribute decision making,then fuzzy clustering is performed through the decision-making score matrix,and the effectiveness of result is validated.Finally,examples prove that the method is feasible and effective.