The invention discloses an iris feature extraction and matching method. The method comprises the steps of performing block segmentation and denoising processing on a standard iris image, and performing iris feature extraction on all denoised sub-blocks to form a feature template; performing one-layer wavelet packet decomposition on each sub-block image, extracting a radial texture feature and an arc texture feature of an iris, and forming a circumferential eigenvector; performing multi-layer wavelet packet decomposition on each sub-block image, extracting a global energy distribution feature and a local energy distribution feature of the iris, and forming an energy eigenvector; forming an iris eigenvector of each sub-block image by the circumferential eigenvector and the energy eigenvector of each sub-block, and forming the feature template of the iris image by the iris eigenvectors of the sub-block images; and performing matching identification on feature templates of two iris images, and gradually performing judgment classification based on similarity and a Hamming distance. According to the scheme, the real-time property and adaptability of iris feature extraction and matching can be effectively improved on the basis of ensuring identification precision.