Currently, our national banks usually use a hierarchical approach based on a point system in credit rating for commercial corporations. Although there are great improvements on rating approach in domestic study, the fuzzy nature of the credit-rating processes may be neglected. This paper presents a fuzzy credit-rating approach: first establish a hierarchical decision structure model and then employ fuzzy integral to aggregate the credit information in a bottom-up way. This technique considers not only the objective evidence but also the relative importance of each criterion. Furthermore, it uses fuzzy sets to describe the criteria, so that the final credit-rating results can reveal the changes of credit information. The five rating levels that describe the final evaluation results can provide more valuable information.