Digital and analog contents, generated in additive manufacturing (AM) processes, may be illegally modified, distributed, and reproduced. In this article, we propose a watermarking scheme to enhance the security of AM. Compared with conventional watermarking methods, our algorithm possesses the following advantages. First, it protects geometric models and printed parts as well as G-code programs. Secondly, it embeds watermarks into both polygonal and volumetric models. Thirdly, our method is capable of creating watermarks inside the interiors and on the surfaces of complex models. Fourth, the watermarks may appear in various forms, including character strings, cavities, embossed bumps, and engraved textures. The proposed watermarking method is composed of the following steps. At first, the input geometric model is converted into a distance field. Then, the watermark is inserted into a region of interest by using self-organizing mapping. Finally, the watermarked model is converted into a G-code program by using a specialized slicer. Several robust methods are also developed to authenticate digital models, G-code programs, and physical parts. These methods perform virtual manufacturing, volume rendering, and image processing to extract watermarks from these contents at first. Then, they employ similarity evaluation and visual comparison to verify the extracted signatures. Some experiments had been conducted to validify the proposed watermarking method. The test results, analysis, discussion, and comparisons are also presented in this article.