Previous meta-analyses investigating attentional biases towards pain have utilized reaction time measures. Eye-tracking methods have been adopted to more directly and reliably assess biases, but this literature has not been synthesized in relation to pain. This meta-analysis aimed to investigate the nature and time-course of attentional biases to pain-related stimuli in participants of all ages with and without chronic pain using eye-tracking studies; and determine the role of task parameters and theoretically relevant moderators. After screening, 24 studies were included with a total sample of 1425 participants. Between-group analyses revealed no significant overall group differences for people with and without chronic pain on biases to pain-related stimuli. Results indicated significant attentional biases towards pain related words or pictures across both groups on probability of first fixation (k = 21, g = 0.43, 95% CI 0.15: 0.71, p = 0.002), how long participants looked at each picture in the first 500ms (500ms epoch dwell: k = 5, g = 0.69, 95% CI 0.034: 1.35, p = 0.039) and how long participants looked at each picture overall (total dwell time: k = 25, g = 0.44, 95% CI 0.15: 0.72, p = 0.003). Follow-up analyses revealed significant attentional biases on probability of first fixation, latency to first fixation and dwell time for facial stimuli, and number of fixations for sensory word stimuli. Moderator analyses revealed substantial influence of task parameters and some influence of threat status and study quality. Findings support biases in both vigilance and attentional maintenance for pain-related stimuli but suggest attentional biases towards pain are ubiquitous and not related to pain status.