Weng et al.introduced the notion of conditional proxy re-encryption(C-PRE),whereby only the ciphertext satisfying one condition set by Alice can be re-encrypted by the proxy.To achieve more fine-grained delegation on the condition set,we introduce a new cryptographic primitive called fuzzy conditional proxy re-encryption(FC-PRE) in which we view conditions in C-PRE as a set of descriptive keywords.A FC-PRE scheme allows for a proxy with the re-encryption key for a conditional keyword set,W,to re-encrypt a ciphertext encrypted with a conditional keyword set,W,if and only if W and W are close to each other as measured by the “set overlap” distance metric.Therefore,our system allows a certain amount of error-tolerance in the conditions.We formalize the FC-PRE security model by incorporating the advantages in previous C-PRE.Finally,we present an efficient construction of FC-PRE scheme,and prove its CCA-security under the wellstudied decisional bilinear Diffie-Hellman(DBDH) assumption in the random oracle model. Weng et al. Introduced the notion of conditional proxy re-encryption (C-PRE), whereby only the ciphertext satisfies one condition set by Alice can be re-encrypted by the proxy. To achieve more fine-grained delegation on the condition set, we introduce a new cryptographic primitive called fuzzy conditional proxy re-encryption (FC-PRE) in which we we view conditions in C-PRE as a set of descriptive keywords. A FC-PRE scheme allows for a proxy with the re-encryption key for a conditional keyword set, W, to re-encrypt a ciphertext encrypted with a conditional keyword set, W, if and only if W and W are close to each other as measured by the “set overlap” distance metric.Therefore, our system allows a certain amount of error-tolerance in the conditions. We formalize the FC-PRE security model by incorporating the advantages in previous C-PRE. Finally, we present an efficient construction of FC-PRE scheme, and prove its CCA-security under the well-established decisional bilinear Diffie-Hellman (DBDH) assumption in the random oracle model