BACKGROUND: According to current fear-avoidance models, changes in motor behaviour (e.g. avoidance) are a key component in the development and maintenance of chronic pain complaints. Yet, experimental research assessing actual behavioural changes following painful events is relatively sparse. This study investigated the effects of pain anticipation on changes in motor behaviour using a fear conditioning paradigm and robot-generated standardized movement trajectories of the upper extremities. METHODS: Pain-free participants (N = 20) performed clockwise and counterclockwise fixed, circular movements with a robotic arm without receiving visual feedback. During fear acquisition, moving in one direction (CS+) was paired with a painful stimulus (pain-US) whereas moving in the other direction (CS-) was not. During the subsequent extinction phase, the pain-US was omitted. We assessed self-reported pain-related fear and urge to avoid the movement, as well as several behavioural measures: Velocity, acceleration, exerted force and force direction. RESULTS: Movements that were paired with pain were associated with increased self-reported pain-related fear and urge to avoid. Moreover, movements that were associated with pain were performed faster, more forcefully and more accurately than movements that were not associated with pain. All these differences diminished during the extinction phase. CONCLUSIONS: The present study demonstrates the utility of robot-generated force feedback in the study of pain-related fear and associated changes in motor behaviour. SIGNIFICANCE: Fear of pain changes movement: Movements associated with pain are performed faster, with more force and higher accuracy than movements that are not associated with pain. These changes can inform us how fear of pain translates into avoidance and escape behaviour, two important constructs in the maintenance of chronic pain.