The standard Unified Modeling Language (UML) and general tools for self-adaptive software could not model Software Fuzzy Self-Adaptation (SFSA) directly in the analysis and design phases. A new approach called Fuzzy Case was proposed by extending UML use case for the modeling of SFSA. By combining the conceptual framework of SFSA and applying the UML extending mechanism, new stereotypes and tagged values were introduced and a meta-model for Fuzzy Case was created. Then, the syntax structure was given and the semantics of Fuzzy Case were defined with Object Constraint Language (OCL). The instance validation shows that, compared with the traditional approaches, Fuzzy Case can express structures of SFSA more explicitly, define semantics of SFSA more accurately and model SFSA more conveniently. The proposed approach plays an important role in improving the development efficiency of SFSA.