In recent years, there were many studies about quality metrics or look-at-point for multi-viewpoint 3D images. Since there are much viewpoints for multi-viewpoint 3D images, we need to consider about digital watermarking, in addition to the coded degradation in multi-viewpoint 3D images. In this paper, at first, we encoded and decoded multi-viewpoint 3D CG images by using H.265/HEVC. Next, we processed digital watermarking for each 3D images after subband division by Haar Wavelet Transformation. Then, we experimented subjective quality evaluation for 3D images including the coded degradation by regions with 8 viewpoints parallax barrier method. Finally, we discussed by analyzing statistically by using Sequential Minimal Optimization (SMO) algorithm of Support Vector Machine (SVM).