An image pickup apparatus which is capable of accurately acquiring evaluation values at target positions for contrast evaluation value acquisition to thereby suppress degradation of accuracy of autofocus due to skipping the reading of a contrast evaluation value at an in-focus position and an increase in autofocus time. The image pickup apparatus acquires from an interchangeable lens at least one of a drivable defocus amount of the interchangeable lens, and a driving speed in driving of a focus lens to focus adjustment positions. The image pickup apparatus changes a lens driving method for focus detection, based on at least one of the drivable defocus amount, the driving speed, and a charge storage interval for charge storage in an image pickup element for acquiring a signal for use in focus detection.