In digital-to-analog/analog-to-digital conversion, printed images mainly involve three color spaces:RGB, CIELAB and CMYK. The work aims to embed watermarking images in different color spaces based on the DCT-domain watermarking algorithm. Six types of images including dark tone, saturated tone, landscape, bright tone and human figure and neutral grey were selected to evaluate the transparency and robustness in each component image in RGB, LAB and CMYK color spaces with three complexity levels respectively. According to the transparency analysis, the transparency of watermark embedded in Lab space b channel, RGB space B channel and CMYK space M channel was better. According to the robustness analysis, images in bright tone had poor resistance to non-geometric attacks in the RGB space. The anti-attack performance of all images in the first three channels of CMYK was better than that in K channel. The anti-attack performance of all images in dark tone was poor in anti-print attack capability. The anti-print performance of all images in the LAB space was the best, among which that in L channel was the best. In this work, the effect of color spaces on the watermarking performance of printed images is systematically studied. The results show that color spaces have different effects on watermarking performance of image of different types and complexity. It improves the theoretical system of digital watermarking technology in the application of printed images, and will provide some guidance and reference for the industrial application and optimization of color image digital watermarking technology.