BACKGROUND: We aimed to quantify the association between chronic musculoskeletal pain and all-cause mortality, and to investigate the extent to which this association is mediated by physical activity, smoking status, alcohol consumption, and opioid use. METHODS: For this population-based cohort study, we used data from UK Biobank, UK between baseline visit (2006-2010) to 18th December 2020. We assessed the associations between chronic musculoskeletal pain and all-cause mortality using a Cox proportional hazards model. We performed causal mediation analyses to examine the proportion of the association between chronic musculoskeletal pain and all-cause mortality. FINDINGS: Of the 384,367 included participants, a total of 187,269 participants reported chronic musculoskeletal pain. Higher number of pain sites was associated with increased risk of all-cause mortality compared to having no pain (e.g., four sites vs no site of pain, Hazard Ratio [HR] 1.46, 95% Confidence Interval [CI] 1.35 to 1.57). The multiple mediator analyses showed that the mediating proportions of all four mediators ranged from 53.4% to 122.6%: among participants with two or more pain sites, the effect estimate reduced substantially, for example, HR reduced from 1.25 (95% CI: 1.21 to 1.30; two pain sites) to 1.07 (95% CI: 1.01 to 1.11; two pain sites). INTERPRETATION: We found that higher number of pain sites was associated with increased risk of all-cause mortality compared to having no pain, and at least half of the association of chronic musculoskeletal pain with increased all-cause mortality may be accounted for by four mediators. FUNDING: Twins Research Australia.