Biometric systems are systems that enable individuals to be recognized in electronic environment using some physical and behavioral characteristics. Iris recognition system is one of the effective biometric recognition systems. The main goal of this study is recognition of the human from the iris images according to the local texture structures. The digital iris images were derived from CASIA database. The texture features were extracted from the four local iris regions of segmented image by using Gray Level CoOccurrence Matrix (GLCM). Totally 88 parameters were extracted for each image as a feature vector. Then, the obtained feature vectors were classified by using k-Nearest Neighbor (k-NN) classifier and the average performance of each system were compared according to different k values (1, 3, 5, 7 and 9). Finally, the best average performance among system architectures of iris recognition system was observed as 85 % in k=1 neighbor structure of k-NN classifier