Abstract Purpose To investigate the longitudinal associations between pain and falls risks in adults. Methods Prospective cohort study on data from 40,636 community-dwelling adults ≥ 50 years assessed in Wave 5 and 6 in the Survey of Health, Ageing and Retirement in Europe (SHARE). Socio-demographic and clinical information was collected at baseline (Wave 5). At 2-year follow-up (Wave 6), falls in the previous 6 months were recorded. The longitudinal associations between pain intensity, number of pain sites and pain in specific anatomic sites, respectively, and falls risk were analysed by binary logistic regression models; odds ratios (95% confidence intervals) were calculated. All analyses were adjusted for socio-demographic and clinical factors and stratified by sex. Results Mean age was 65.8 years (standard deviation 9.3; range 50–103); 22,486 (55.3%) participants were women. At follow-up, 2805 (6.9%) participants reported fall(s) in the previous 6 months. After adjustment, participants with moderate and severe pain at baseline had an increased falls risk at follow-up of 1.35 (1.21–1.51) and 1.52 (1.31–1.75), respectively, compared to those without pain (both p < 0.001); mild pain was not associated with falls risk. Associations between pain intensity and falls risk were greater at younger age (p for interaction < 0.001). Among participants with pain, pain in ≥ 2 sites or all over (multisite pain) was associated with an increased falls risk of 1.29 (1.14–1.45) compared to pain in one site (p < 0.001). Conclusions Moderate, severe and multisite pain were associated with an increased risk of subsequent falls in adults.