An image capturing apparatus is provided with an image sensor having phase difference detection pixels for performing focus detection, an exposure control unit for causing at least the phase difference detection pixels to perform a plurality of accumulations whose exposure periods are a second exposure period shorter than a first exposure period which is a target for obtaining a main image by the image sensor, and a prediction unit for predicting an in-focus position of a subject based on signals of at least two phase difference detection pixels obtained by at least two accumulations among the plurality of accumulations whose exposure periods are the second exposure period.