Estimating software effort and costs is a very important activity that includes very uncertain elements. The concepts of the Fuzzy Set Theory has been successfully used for extending metrics such as FP and reducing human influence in the estimation process. However, when we consider object-oriented technologies, other models, such as the Use Case model, are used to represent the specification in the early stages of development. New metrics based on this model were proposed and the application of the Fuzzy Set Theory in this context is also very important. This work introduces the metric FUSP (Fuzzy Use Case Size Points) that allows gradual classifications in the estimation by using fuzzy numbers. Results of a study case show some advantages and limitations of the proposed metric.