<p>Humans control their body movements by exploiting gravity to minimise muscle effort while achieving task goals. Most of these findings have been observed in physical environments, although some have also been confirmed in virtual environments. However, research using virtual environments to explore gravity-related motor control mechanisms has yet to directly compare motor performance between virtual and physical environments. Therefore, the present study aimed to examine in detail the potential differences in upper-limb pointing movements between virtual and physical environments. To this end, participants performed pointing tasks in four directions (upward, downward, leftward, and rightward, from an allocentric perspective) in both upright and lying postures, under both virtual and physical conditions. Our results showed that relative duration to peak velocity—a well-established kinematic indicator of gravity utilisation—was consistently shorter for upward than for downward movements across both environments and both postures. However, no differences were observed between the two environments when posture and movement direction were held constant. Furthermore, no differences were observed between the environments in terms of whole velocity and acceleration profiles, as well as in movement duration, peak velocity, peak acceleration, peak deceleration, and the relative durations to peak acceleration and peak deceleration. The similarity in relative duration to peak velocity between virtual and physical environments suggests that the effects of gravity on pointing movements can be reliably assessed in virtual environments as in physical ones. This supports the use of virtual environments as valid tools for studying pointing movements.</p>