An image forming apparatus has an image carrier on which an electrostatic latent image is formed, a developing unit that develops the electrostatic latent image using a toner to form a toner image, and a cleaning unit that removes the toner remaining on the image carrier after transferring the toner image to a recording medium. The image forming apparatus includes a buffer housing unit that temporarily houses the toner removed by the cleaning unit, a first carrier unit comprising a carrier driving member, the first carrier unit provided between the cleaning unit and the buffer housing unit, and the first carrier unit carries the toner removed by the cleaning unit to the buffer housing unit using the a carrier driving member, a second carrier unit comprising a carrier driving member, the second carrier unit provided between the buffer housing unit and the developing unit, and the second carrier unit carries the toner housed in the buffer housing unit to the developing unit using the carrier driving member, a waste housing unit connected to the buffer housing unit, and a control unit that separately controls the carrier driving member of the first carrier unit and the carrier driving member of the second carrier unit, wherein the buffer housing unit has an upper limit of housing capacity, and the waste housing unit houses the toner exceeding the upper limit and overflowing from the buffer unit.