An image forming apparatus which reduces downtime by starting image density detection without waiting for identification of the same position on a rotary member. A detection unit detects a surface condition of a rotary member on which a toner image on a photosensitive drum is transferred. An identifying unit matches first data detected by the detection unit on a reference point on the surface of the rotary member and second data detected by the detection unit after obtainment of the first data to identify the reference point. A control unit controls image density based on a detection result obtained by the detection unit on a base of the rotary member corresponding to a range where a measurement image is formed and an obtained detection result on the measurement image by referring to the reference point identified as a reference.