A vehicle driving assist system includes an imaging device, an alarm, and a controller. The controller includes a facial recognition unit, an occupant determination unit, and a loss determination unit. The facial recognition unit recognizes at least a part of an occupant's face in first captured images acquired from the imaging device. The occupant determination unit determines a driving state of the occupant on the basis of a second captured image in which at least the part of the face is recognizable, and generate a first alert request if the occupant is unable to drive the vehicle. The loss determination unit generates a second alert request if a third captured image in which at least the part of the face is unrecognizable occurs two or more times. The alarm outputs an alert to the occupant on the basis of the first and second alert requests.