Introduction This study aims to assess the reliability of the Danish version of the Canadian Occupational Performance Measure (COPM). Method 151 clients, 42.4% male, mean age 66.9 years ( SD: 14.4, range: 16–90), from two hospitals and two rehabilitation centres were interviewed twice with the COPM over a ten-day mean interval ( SD: 4.9, range: 2–27) either by the same (intra-rater) or two different occupational therapists (inter-rater). Data were analysed with intra-class correlation coefficients, coefficient of repeatability, and Bland–Altman plots. Results 823 occupational performance issues were prioritized of which 41%, 338 occupational performance issues (95% CI: 37.7–44.5), were mentioned in both interviews. The intra-class correlation coefficients were 59.8 (95% CI-intra-class correlation coefficients: 49.3–69.5) (COPM), 73.5 (95% CI intra-class correlation coefficients: 65.5-80.2) (COPM-Performance), and 71.8 (95% CI intra-class correlation coefficients: 63.4–78.8) (COPM-Satisfaction). The limits of agreement were −2.83 to 3.05 for the COPM-Performance and −3.50–3.53 for the COPM-Satisfaction. The COPM-Importance scores were highest when obtained by two raters, but opposite for the scores of the COPM-Performance and COPM-Satisfaction. The coefficient of repeatability roughly showed a change in +/- three points (COPM-Importance: 2.67; COPM-Performance: 2.94; and COPM-Satisfaction: 3.52). Correspondingly, Bland–Altman plots showed limits of agreement for the mean values of −2.71 and 3.53, respectively. Conclusions The intra-rater and inter-rater reliability of the COPM were moderate across settings, clients and rater experiences.