Abstract: The fact that lots of color JPEG images are manipulated and altered has made it more and more important to identify the color image forgeries. Many image detection methods are proposed for gray and uncompressed images and most of them can’t identify these compressed color forgeries effectively. In reaction to the phenomenon this paper presents a passive-blind approach to detect spliced color image forgeries based on distortion properties after double JPEG quantization. By analyzing the process of manufacturing forgeries, we got the point that different regions in high-quality color forgeries which are composed from different JPEG images have different distortion due to different double quantization process. So we compress the forgeries again using the estimated initial quantization tables of the background area firstly. Then, we define the distortion function for each color-component in the recompressed images. By analyzing the value of each function in different areas in the processed images, we can confirm different tampered areas corresponding to different color components. We synthesize the testing results of all the color components and confirm the dimension and the position of the tampered area. The simulation results show that the approach is very effective in detecting the tampered area of the color image forgeries and is better than the way based solely on a color component.