In recent years, digital image watermarking has attracted much attention from researchers because of its great ability to ensure the security and protection of information against unauthorized access or illegal modification. In this paper, a robust watermarking scheme for medical images based on the discrete wavelet transform (DWT), the singular value decomposition (SVD), and the Arnold transform (AT) has been proposed. Several image modalities including ultrasound, CT scan, PET scan, X-ray, and MRI were used to test our proposed method. On the other hand, two metrics PSNR and SSIM were used to assess the imperceptibility while the NC metric was used to evaluate the robustness. Experimental results without attacks showed that our method offers good imperceptibility with a PSNR above 45dB and SSIM close to 1. After applying the attacks, the results showed that the method presents good robustness with NC close to 1. In conclusion, our proposed method offers a good compromise between imperceptibility and robustness which allows it to be a valuable method in the field of e-health.