An augmented reality-based video calling system includes a first terminal configured to capture a video and a second terminal configured to add virtual content to the video captured by the first terminal and provide information regarding the added virtual content to the first terminal. The first terminal calculates 3D position coordinates of the virtual content in a 3D space corresponding to the video on the basis of the provided information regarding the virtual content, combines the virtual content with the video using the calculated 3D position coordinates, and display the virtual content combined with the video.