An output shaft supporting structure includes an output shaft having a hollow space opened in one end side, an intermediate shaft whose one end having an insert portion inserted into the hollow space, an extension housing placed on the edge of a casing, a first bearing and a second bearing disposed between the outer peripheral surface of the output shaft and the inner peripheral surface of the extension housing, a third bearing disposed between the outer peripheral surface of the insert portion and the inner peripheral surface of the hollow space. A part of rotational elements of a planetary gear set is supported between the outer peripheral surface of the insert portion closer to one end of the intermediate shaft than the third bearing and the inner peripheral surface of the hollow space, and at least a portion of axial position of the third bearing is located between the first bearing and the second bearing.