Metacognitions, which are beliefs about our own thinking processes, can modulate worry and rumination and thereby influence emotional distress. This study aimed to develop a self-report measure of unhelpful pain-related metacognitions which might serve as a clinical and research tool to better understand pain catastrophizing, a significant risk factor for adverse pain outcomes. Two phases of validation are presented. Phase 1 reports on how the Pain Metacognitions Questionnaire (PMQ) was empirically developed through a qualitative study of 20 people with chronic back ( = 15) or knee ( = 5) pain in secondary or tertiary care and then validated in a large internet sample of people experiencing pain ( = 864). Rasch analysis yielded a 21-item scale with two dimensions (positive and negative metacognition) assessing how useful and problematic people believe rumination about pain to be, respectively. In Phase 2, further validation using a new sample ( = 510) replicated initial findings. Both PMQ subscales have good retest reliability ( = 0.76, = 0.72) and internal consistency (0.86, 0.87). They correlate negatively with mindfulness and positively with pain intensity, disability, anxiety, depression, catastrophizing, rumination, and metacognition. The PMQ also predicts unique variance in catastrophizing when other variables are controlled and predicts 'patient' status for pain catastrophizing. Sensitivity analysis yielded preliminary suggestions for clinically meaningful cut-offs. Unhelpful pain metacognitions can be validly and reliably measured using a self-report instrument. Future studies using the PMQ might shed new light on pain-related thinking processes to develop better interventions for people prone to worry and rumination about their pain.