ing at the problems of the insufficient sampling and sensitivity to random noise and non-uniform illumination of the local binary pattern, a face recognition method of the improved gradient local binary pattern (IGLBP) is proposed. Two groups of 3 pixel×3 pixel subneighborhood are obtained by the multi-radius and multi-direction sampling mode, including 16 pixels in two radii and eight directions. The features are extracted by the gradient local binary pattern, and then the two sets of them are encoded to produce IGLBP. Finally, the IGLBP feature is divided to get the feature vector of the face according to the block histogram, and it is used for classification and recognition. The experimental results of CAS-PEAL and AR face database show that the proposed algorithm can effectively extract the feature information, and it is robust to variations of the illumination, expression, partial occlusion and noise in face recognition.