Color and color neighborhood statistics have been used extensively in image matching and retrieval. However, the effective incorporation of color layout information remains a challenging issue. In this paper we present a novel method for color layout based image matching called Spatial Color Component Matching (SCCM). First perceptually dominant colors are extracted from an image and are back-projected to segment the image into various areas. Then, each dominant color area, depending on its size, is segmented into a number of spatial units using a multilevel graph partitioning algorithm. Each unit is described in terms of its color and a set of spatial attributes to form a Spatial Color Component (SCC). All SCC?s form a list that summarizes the color layout information in an image. The distance between two images is then defined by the minimum distance mapping between the two corresponding SCC lists. The algorithm has been evaluated using an image database of wall paper patterns and another database of natural images. It has been judged by human subjects to be highly effective in both cases.