An image forming apparatus includes: a primary transfer section which executes a first transfer by superimposing respective color toner images formed on a plurality of photoreceptors onto an intermediate transfer member; a secondary transfer section including a transfer roller, which concurrently transfers a plurality of toner images superimposed and formed on the intermediate transfer member onto a sheet conveyed from a sheet feed tray, which is in contact with the intermediate transfer member, and a high voltage power source which applies voltage to the transfer roller; and a controller which judges whether there is residual toner, which is a toner image formed onto the intermediate transfer member but not transferred onto the sheet when a sheet sensor detects the sheet exhaustion, and executes a cleaning mode in which the transfer roller is cleaned when there is the residual toner.