An image pickup apparatus enabling release time lag reduction. A first holder holds a first mirror and is movable between a first position in an optical path and a second position retracted from the optical path. A second holder holding a second mirror in a state rotatably attached to the first holder is movable between a third position in the optical path and a fourth position retracted from the optical path. An MPU causes an mirror drive unit to rotate the second holder in the third position toward the first holder in the first position, after termination of focus detection when the first and second holder are in the first and third positions, respectively, and before image pickup is instructed, to thereby move the second holder to a fifth position closer to the first holder than the third position is.