An image forming apparatus comprises: an intermediate transfer belt having a belt surface on which a toner image is formed while the belt surface is moved and transferring the toner image that has been formed on the belt surface onto a sheet to be fed by a feed section; a resist roller pair that is disposed at a position on the upstream side from a transfer position at which the toner image that has been formed on the belt surface is transferred onto the sheet in the sheet feeding direction and performs a resist operation before the sheet has been fed to the transfer position as well as constitutes the feed section; a belt surface sensor that is disposed at a position on the downstream side from the transfer position in the belt surface moving direction such that the sheet moving distance from the sheet holding position between the resist roller pair to the position becomes shorter than the length of the sheet in the feeding direction and detects the sheet on the belt surface; and a control unit that allows the feed section to stop the sheet feeding operation at the time when the belt surface sensor detects the sheet.