A stereoscopic image processing method capable of preventing occurrence of undesirable cutout of image when a stereoscopic combined image is printed out is provided.A stereoscopic combined image can be created and recorded using a digital still camera 1. This stereoscopic combined image is a single image formed by arranging a plurality of view-point images. At first, left view-point image data and right view-point image data are stored in a memory by performing photographing twice. Then, the data of the two images are arranged laterally into a single image. Through this processing, original stereoscopic combined image data is created. Padding areas (dummy areas) are added to the original stereoscopic combined image data at the upper and lower sides thereof. Namely, a processed image containing the original stereoscopic combined image is created.