The contents-based image retrieval (CBIR) is general type of retrieval which has been an active area of research for many years. How to quickly retrieval the image is a very action research topic in the research of image retrieval based on Web because of the large amount of data required by images. Therefore automatic and efficient indexing is needed for fast content based retrieval, it alleviates the drawback of any manual annotating. The main focus of this study is dimensionality reduction and image index of Web image. First, the paper presents the commonly used methods for image index. Then, it describes how to convert from RGB model to HSV model, and how to extract 72-dimensional feature of image based on HSV color space. In the end, the method about 72-trees for image index is discussed. The experiments and results, which based on Corel database, showed that this method has greatly improved the image retrieval in time and precision rates.