The primary aim of the current study was to experimentally test whether pain-related fear can be acquired through observational learning, whether extinction occurs after actual exposure to the aversive stimulus, and whether pain-related fear was associated with increased pain ratings. During an observation phase, female volunteers watched a video showing models performing cold pressor tasks (CPT), of which the color served as a conditioned stimulus (CS). In a differential fear conditioning paradigm, each of two colors were either paired with models’ painful (CS+) or neutral (CS-) facial expressions. Exposure consisted of participants performing CPTs of both colors (10°C). Self-reported fear of pain, and expected pain ratings were obtained after the observation period, while actual pain and avoidance measures were obtained during and after exposure. Results show that after observing another person performing the CPT associated with the painful faces, subjects report more fear of pain and expect more intense and unpleasant pain as compared to the CPT associated with the neutral faces. This effect of observational learning on pain-related fear persisted until after exposure. During and after exposure no stimulus type effect for pain ratings was found. This study provides preliminary evidence for observational learning of pain-related fear in humans.