A method and an image pick-up apparatus for calculating coordinates of an object being captured by using fisheye images are provided. In the method, a first and a second fisheye image containing the object are respectively captured by using a first and a second fisheye lens. Coordinates of the object in the first and the second fisheye image are detected. A first and a second azimuth angle of the object relative to fisheye centers of the first and the second fisheye image on an image sensor plane are accordingly calculated. Distances between the coordinates of the object and the fisheye centers are respectively transformed into a first and a second incident angle through lens curves of the fisheye lenses. Three-dimensional coordinates of the object are calculated according to the first and the second azimuth angle, the first and the second incident angle, and a baseline distance between the fisheye lenses.