To describe sociodemographic disparities in temporal trends of incidence and age distributions of first registered osteoarthritis (OA) diagnosis in southern Sweden. We identified all Skåne residents aged 35+ who had lived in the region at any point during the period 2006-2019 with no previous OA diagnosis (ICD-10 codes M15-M19) for 8 years prior to inclusion in the study (n = 849,061). We calculated person-years from inclusion until OA diagnosis, death, emigration, or 31 December 2019, whichever occurred first. Combining sex (female, male), education (low, medium, high) and nativity (Swedish, immigrant), we created a variable with 12 strata. Average annual percent changes in age-standardized incidence rates were estimated using joinpoint regression. Changes in the median age-at-diagnosis (year of diagnosis minus birth year), weighted to the mid-2005 Swedish population, were explored. Cumulative age-standardized incidence rates ranged from 116 (95% CI: 111, 121) per 10,000 person-years for immigrant males with low education to 205 (95% CI: 200, 210) for immigrant females with medium education. The estimated average annual percent changes (ranging from 3.4% to 6.1%) were generally similar, with slightly greater variations among immigrants than Swedes. The weighted median age-at-diagnosis was higher for Swedes and low educated people. Immigrant females with low education were the only stratum with a reduction (3 years) in the weighted median age-at-diagnosis over time. Sociodemographic patterns in knee OA incidence were different from patterns for hip OA. There were few sociodemographic disparities in temporal trends of OA incidence and age-at-diagnosis, suggesting persistent sociodemographic disparities in OA burden in southern Sweden.