Personal identification system using biometric features of individuals, many applications in different industries are secure. These systems typically focus on an individual feature and its detection is done by processing the input images. In this paper, we proposed a new method for the synthesis of useful features that we used to identify a person . First, the image sensor fingerprint and iris features on offer and then they are combined . Next, processing is performed by a trained neural network . In order to compare our proposed method with other existing methods, a standard data source is used. The source characteristics of fingerprints and iris images , there are separate persons . Based on the fitness values , runtime error rate of the system, it was found that the composition and processing characteristics using the proposed algorithm is much better than either one alone .