A method for capturing a three dimensional (3D) image by using a single-lens camera is provided. First, a first image is captured. According to a focus distance of the single-lens camera in capturing the first image and an average distance between two human eyes, an overlap width between the first image and a second image required for capturing the second image of the 3D image is calculated. Then, the first image and a real-time image captured by the single-lens camera are displayed, and an overlap area is marked on the first image according to the calculated overlap width. A horizontal shift of the single-lens camera is adjusted, to locate the real-time image in the overlap area. Finally, the real-time image is captured as the second image, and the first and second images are output as the 3D image.