BACKGROUND: Pregnant women are susceptible to ambient temperature which may associate with adverse birth outcomes. These associations might change over time due to adaptation. However, no evidence is available worldwide. OBJECTIVES: To examine the effects of ambient temperature at three trimesters of pregnancy on preterm birth and stillbirth, and evaluate the effect changes during 1994-2013. METHODS: Information on birth outcomes and meteorological parameters during 1994-2013 were obtained in Brisbane, Australia. Cox proportional hazards models were employed to evaluate the effects of mean temperature during each trimester of pregnancy on preterm birth and stillbirth. An interaction term between time and temperature was used to assess the potential changes in effects of temperature over time. RESULTS: Both low and high temperatures at the 3rd trimester of pregnancy significantly increased the risk of preterm birth, with similar hazard ratios (HRs) [95% confidence intervals (CIs)] for low [1.21 (1.16, 1.27)] and high [1.21 (1.16, 1.26)] temperatures in comparison with thresholds. Increased risk of stillbirth was significantly associated with both low and high temperatures at the 2nd trimester of pregnancy, and lower HRs were observed for low temperature [1.23 (1.04, 1.45)] than high temperature [1.47 (1.24, 1.74)], in comparison with thresholds. The effects of low temperature became stronger, whereas the effects of high temperature became weaker from 1994 to 2013 for both preterm birth and stillbirth. CONCLUSIONS: Both low and high ambient temperatures during pregnancy increase the risks of preterm birth and stillbirth. Pregnant women should have adapted to high temperature but not low temperature.