Regular stress checks and health confirmations by supervisors are standard practices in high-risk environments such as construction sites. Therefore, effective health management of onsite workers is crucial for preventing health hazards and casualties at the workplace. Nevertheless, detection of falsified health declarations and unrecognized health issues remains a challenge. This study used facial thermography obtained by infrared thermal imaging to monitor the health conditions of workers. Thermal images of the face reflect the blood flow distribution in the skin, which is controlled by the autonomic nervous system and can serve as noninvasive health status indicators. However, ambient temperature and time-of-day variations affect skin temperature readings, making the application of facial thermal imaging challenging in uncontrolled environments outside laboratory setting. To address this issue, we proposed a method to correct the artifacts caused by environmental temperature and time-of-day variations in facial thermography. This method provides standard skin temperatures for any given environmental temperature and time of the day, allowing comparison with actual skin temperatures. We conducted a five-month longitudinal study in 279 construction workers, and collected 2,884 pairs of facial thermal images and health status data. The collected face thermal images were corrected for environmental temperature and time-of-day variations. The effectiveness of this method in assessing the health status of onsite workers was evaluated. The results showed that the proposed method effectively corrected artifacts from environmental temperature and time-of-day variations in facial thermography. Moreover, the corrected images revealed that a decline in health conditions was associated with a decrease in temperature around the nose and an increase in temperature in other areas. This is consistent with previous findings of the relationship between stress and facial skin temperature.