BACKGROUND: Evidence on long-term trends in gestational diabetes mellitus (GDM) prevalence in Australia is lacking. AIMS: To assess and compare trends in GDM prevalence among Indigenous and non-Indigenous Australian women. MATERIALS AND METHODS: Analysis of crude and age-adjusted GDM prevalence over time by Indigenous status and age, using routinely collected midwives data from Australian states and territories on mothers giving birth from 1990 to 2009. RESULTS: Despite considerable data variation, particularly in 1990-1999, and likely underestimation of GDM prevalence, crude and age-adjusted GDM prevalences were higher in Indigenous than non-Indigenous women at all time-points (4.7% vs 3.1% in 1990-1999; 5.1% vs 4.5% in 2000-2009, P < 0.0001). Data variability precluded quantitative assessment of trends and changes in prevalence ratios before 2000. From 2000 to 2009, GDM prevalence increased significantly among Indigenous women by a mean 2.6% annually (Ptrend <0.0001), and non-Indigenous women by 3.2% annually (Ptrend <0.0001), with no significant trend in the age-adjusted Indigenous/non-Indigenous prevalence ratios (PR) (P = 0.34). GDM prevalence increased significantly with age (P < 0.0001), although the increase with age was significantly greater among Indigenous women (PR 5.34 (4.94-5.77), ≥35 vs <25 years) compared to non-Indigenous women (PR 3.72 (3.64-3.81), ≥35 vs <25 years), Pinteraction <0.0001. CONCLUSIONS: Bearing data quality concerns in mind, GDM prevalence is increasing rapidly among Australian women, more than doubling in non-Indigenous women between 1990 and 2009. Prevalence is consistently higher in Indigenous versus non-Indigenous women, with statistically consistent differences between the groups in recent years. The marked increase in prevalence with age highlights an important period for prevention, particularly for Indigenous women.