Biometric cryptography is a technique using biometric features to encrypt the data, which can improve the security of the encrypted data and overcome the shortcomings of the traditional cryptography. This paper proposes a novel biometric cryptographic algorithm based on the most accurate biometric feature -- iris. In this algorithm, a 256-dimension textural feature vector is extracted from the preprocessed iris image by using a set of 2-D Gabor filters. And then the Add/Subtraction operations and Reed-Solomon error-correcting algorithm are employed to directly encrypt and decrypt the data. Experimental results demonstrate the feasibility of the proposed system.