An image forming apparatus includes an image bearing belt configured to bear a toner image, a driving roller configured to rotationally drive the image bearing belt, first and second image forming units configured to form toner images on the image bearing belt. To correct color misregistration, the image forming apparatus uses a mark detecting unit including a plurality of sensors provided at different positions in a rotation direction of the image bearing belt to detect a mark on the image bearing belt, and adjusts a rotational speed of the driving roller, based on detection results of the mark detecting unit and temperature detected within the image forming apparatus.