Biometric identification using modalities based on the human body is being actively studied. For physical biometrics, such as fingerprints, veins, irises, and faces, which have been widely used for identification, spoofing is recognized as a serious vulnerability. In an attempt to prevent spoofing, here, we have proposed three new biometrics approaches based on various features of the eyes, namely, the behavioral characteristics of eye movement, three-dimensional shape of the iris surface, and a combination of measuring the contour of the upper eyelid during blinking, respectively. All three proposed methods have high accuracy of identification and are resistant to spoofing. In this chapter, the respective characteristics, experimental devices, algorithms of image processing, evaluative experiments, and summary of each proposed method are described. In the final section, the characteristics and accuracy of each method are compared and discussed.