Abstract: This paper proposes an image retrieval algorithm based on global and local color features.First, global color features are extracted from Lab color space which is consistent with human visual perception, and Gaussian weighting coefficients are applied to the image which is divided into sub-blocks.Then the binary color bitmap is used as local color feature.Furthermore, directions are taken into account and the image is horizontally and vertically projected.At last, similarities of global and local color features are combined reasonably in image retrieval.The results using Corel Image Database show that the proposed algorithm obviously outperforms previous work1.