The vast improvement of electronic commerce paved the way to the fear of terrorism since the traditional ways of personal identification like ID cards and passwords are no longer sufficient. Biometrics is the more secure option that uses parts of a body for authentication and thus is practically impossible to get lost, stolen or forgotten. Biometric identification is an emerging technology that can solve security problem in our worked society. Biometric traits such as fingerprints, hand geometric, face and voice verification and iris recognition provide a reliable alternative for identify verification/identification and are gaining commercial and high user acceptability rate. The iris is one of the most useful traits for biometric recognition. This thesis presents proposed recognition system based on semantic indexing technique. The generic structure of the proposed iris recognition system is presented in details, it is built depending on assuming that the good authentication system should be characterized by using features that are; (i) Highly unique; the chance of any two people having same characteristics will be minimal, and (ii) Stable; the feature does not change over time, and be easily captured in order to provide convenience to the user, and prevent misrepresentation of the feature. The proposed system consists of two phases: the collection and recognition. The collection phase process of storing the characteristics of all models of iris images, while the recognition phase process on the characteristics of the iris compared to the input image with those stored in the database and issue a resolution of recognition depending on the results of the comparison. Characteristics were used Singular Value Decomposition (SVD) and Upper Vector (U) has been found that they share the same amount of recognition in the decision making was therefore possible to use only one to give the same results of recognition. The results of recognition by aggregating Results have been encouraging and it longer opportunity of development and this confirms the correctness of choice of technique used. The recognition rate was more than 98%. The proposed system first find the iris localization by detect the center of pupil and pupil circle. Then find the iris detection by find the center of iris and iris circle. Then make mask image and compare it with eye image in order to separate the iris from eye image. Then convert the iris into binary code. Then make the decision making in order to either enrollment the iris into lookup table or make the recognition decision on it.