Disclosed are training method, program and computing device to ameliorate visual field defect. The method includes presenting, by a computer, a virtual first object on a screen, wherein the first object is presented in an approaching manner toward the subject from a first position to a second position, when a position of the first object reaches the second position, the dividing, by the computer, of the first object into objects and presenting the divided objects on the screen or removing, by the computer, the first object from the screen, when the first object is divided or removed, presenting, by the computer, a virtual second object and a virtual third object on the screen, and receiving, by the computer, an identification input about the second object and the third object from a response input device of the subject.