An image forming apparatus that transfers a toner image on a photosensitive element to a recording medium includes a transfer unit that includes a transfer member in a form of an endless belt pressed against an outer surface of the photosensitive element, and that is detachable from a main unit of the image forming apparatus; a first cleaning unit that cleans residual toner on the transfer member; and a waste toner receptacle that collects the residual toner cleaned by the first cleaning unit. The waste toner receptacle is arranged in an inner circumference of the transfer member. The waste toner receptacle is held inside the main unit, so that the transfer unit is detachable from the main unit.