A biometric system provides automatic identification of a character based on a unique feature or attribute possessed by the creature. Most practical iris recognition systems use original algorithms developed by Daugman, and these algorithms are competent to create perfect recognition rates but not includes time requirement in account and extremely suffered when pupil is illuminated with light intensity. For the efficient pupil detection a morphological bridged canny edge detection technique is proposed, which is especially designed to handle pupil occlusion problem. For the feature selection and mining part, most conventional iris recognition systems relies on wavelet features extracted from rubber sheet model, but the wavelet feature extraction is a complex and time consuming process, to handle this problem in this paper rubber sheet model of iris part is exclusively used as feature. The third and most crucial modification proposed in this paper, is in the feature classification part, after feature space formation using rubber sheet model, K- Nearest Neighbor classifier is proposed to achieve highest iris recognition efficiency. This paper, proposes three serious modifications in available conventional iris recognition method for real time and efficient iris recognition. Mean wise modification proposed are in the pupil segmentation part, Feature extraction part for conduct pupil illumination problem and iris matching part for providing high speed and efficient iris recognition.