The autonomic nervous system (ANS) regulates physiological changes during pregnancy, supporting fetal development and homeostasis. Heart rate (HR) and heart rate variability (HRV) are well-established, non-invasive biomarkers of ANS function. However, their circadian dynamics during pregnancy remain under-explored due to the lack of continuous data collection, a gap now addressed by wearable technology. This study is the first comprehensive investigation of HR and HRV circadian rhythms throughout pregnancy using wearable devices in a free-living setting. We extract longitudinal HR and HRV from smartwatch Photoplethysmography (PPG) data via a machine learning-based pipeline and employ Cosinor analysis to assess circadian rhythm characteristics. Our findings revealed significant HR and HRV circadian patterns over 16 weeks, showing a decline in HRV, an increase in HR rhythm-adjusted mean, and elevated nighttime stress linked to sleep disturbances and daytime fatigue. These results offer valuable insights into ANS regulation during pregnancy and highlight the potential of HR and HRV circadian rhythm parameters as biomarkers for pregnancy complications.