Data substantiating the optimal patient body temperature during cooling procedures in cardiac operations are currently unavailable. To explore the optimal temperature strategy, we examined the association between temperature management and survival among patients during cardiopulmonary bypass assisted coronary artery bypass grafting (CABG) procedures on 30-days and 5-year postoperative survival. Adult patients (n = 5,672, 23.6% female and mean (SD) age of 66 (10) years) operated between 1997 and 2015 were included, with continuous measured intraoperative nasopharyngeal temperatures. The association between mortality and patient characteristics, laboratory parameters, the lowest intraoperative plateau temperature and intraoperative cooling/rewarming rates were examined by multivariate Cox regression analysis. Machine learning-based cluster analysis was used to identify patient subgroups based on pre-cooling parameters and explore whether specific subgroups benefitted from a particular temperature management. Mild hypothermia (32-35°C) was independently associated with improved 30-days and 5-year survival compared to patients in other temperature categories regardless of operation year. 30 days and 5-year survival were 98% and 88% in the mild hypothermia group, whereas it amounted 93% and 80% in the severe hypothermia (<30°C). Normothermia (35-37°C) showed the lowest survival after 30 days and 5 years amounting 93% and 72%, respectively. Cluster analysis identified 8 distinct patient subgroups principally defined by gender, age, kidney function and weight. The full cohort and all patient subgroups displayed the highest survival at a temperature of 32°C. Given these associations, further prospective randomized controlled trials are needed to ascertain optimal patient temperatures during CPB.