Background: Working longer than the maximum recommended hours is associated with an increased risk of cardiovascular disease, but the relationship of excess working hours with incident cancer is unclear. Methods: This multi-cohort study examined the association between working hours and cancer risk in 116 462 men and women who were free of cancer at baseline. Incident cancers were ascertained from national cancer, hospitalisation and death registers; weekly working hours were self-reported. Results: During median follow-up of 10.8 years, 4371 participants developed cancer (n colorectal cancer: 393; n lung cancer: 247; n breast cancer: 833; and n prostate cancer: 534). We found no clear evidence for an association between working hours and the overall cancer risk. Working hours were also unrelated the risk of incident colorectal, lung or prostate cancers. Working ⩾55 h per week was associated with 1.60-fold (95% confidence interval 1.12–2.29) increase in female breast cancer risk independently of age, socioeconomic position, shift- and night-time work and lifestyle factors, but this observation may have been influenced by residual confounding from parity. Conclusions: Our findings suggest that working long hours is unrelated to the overall cancer risk or the risk of lung, colorectal or prostate cancers. The observed association with breast cancer would warrant further research.