A method for spatial interaction in Augmented Reality (AR) includes displaying an AR scene that includes an image of a real-world scene, a virtual target object, and a virtual cursor. A position of the virtual cursor is provided according to a first coordinate system within the AR scene. A user device tracks a pose of the user device relative to a user hand according to a second coordinate system. The second coordinate system is mapped to the first coordinate system to control movements of the virtual cursor. In a first mapping mode, virtual cursor movement is controlled to change a distance between the virtual cursor and the virtual target object. In a second mapping mode, virtual cursor movement is controlled to manipulate the virtual target object. User input is detected to control which of the first mapping mode or the second mapping mode is used.