There are growing concerns about increasing carsickness in a self-driving car as drivers perform various non-driving tasks during autonomous driving. This study investigated the relations between human head motion and motion sickness in laboratory and field experiments. In both experiments, measurement of head motion and subjective evaluation on motion sickness was carried out. In the laboratory experiment, twelve healthy adult participants sitting in a vehicle seat were exposed to 30 min of ±10-degree, 0.1-Hz sinusoidal roll oscillation in the following four conditions to vary head motion: Normal, with a travel pillow, with a weight at the forehead and with weights on sides of the head. Results showed that differences in head motion affect the degree of motion sickness and suggested that yaw head motion had less effect on motion sickness than roll motion. In the field experiment, eight healthy adult participants sat in a minivan’s second-row seat. They experienced 30 minutes of vehicle journeys in the four conditions in combination with two seating postures (upright and relaxed) and two types of headrests (a standard headrest and a headrest with occipital bone support (OBS) to reduce head roll motion effectively). Results showed that OBS significantly reduced the roll and lateral head motion and reduced the severity of motion sickness by 42 to 51%.