Image retrieval technology can be used to find the color painting similarities related to the project and to make full use of relevant data and images from the architectural paintings database. A retrieval method for architectural color paintings based on shape and color features is set up. The first step is to classify the basic types of architectural color paintings. Then the edges of the color painting image are recorded. Then the detected image is matched by the pre-defined template and the image is classified as to one of the basic types of color paintings according to the matching result. The second step is to calculate the similarity within the same type of the images. The color histograms of the color paintings in HSV space are statistically calculated and the search is carried out by the similarity of the histograms within the same type of the color paintings. Experiments show that the retrieval results reached by the method are similar to that made by the perception of human eyes.