It is difficult to effectively distinguish the similarity between colors in RGB color space, so color image processing is implemented in HSI space which reflects the features of human vision. An adaptive algorithm of color image edge detection based on multiple structure and multi-scale elements in HSI space is proposed. Firstly, morphological edge detection is executed by using different structure and different scale elements to hue, saturation and intensity. Then, according to the weight derived from information entropy, the color edge information is obtained by combining component’s edge information. The experimental results show that proposed algorithm can make full use of the hue, saturation and intensity information to effectively eliminate the noise and adaptively extract the complete edge information.