During autonomous driving of a vehicle, useful Information is presented to an occupant. An image generator in driving assistance device generates a first image representing an action the vehicle is capable of executing depending on a travel environment during autonomous driving and a second image representing a basis of the action. Image output unit outputs the first image and the second image generated by the image generator to notification device in the vehicle in association with each other. The action vehicle is capable of executing may be selected from among action candidates by a driver.