To explore the changes of serum procalcitonin (PCT) level in patients with moderate and severe acute respiratory distress syndrome (ARDS) after cardiac surgery under cardiopulmonary bypass (CPB), and try to find out the best cut-off of PCT to predict the progression to moderate and severe ARDS. Medical records of patients undergoing cardiac surgery with CPB in Fujian Provincial Hospital from January 2017 to December 2019 were retrospectively analyzed. Adult patients who were admitted in intensive care unit (ICU) for more than 1 day and had PCT values on the first postoperative day were enrolled. Clinical data such as patient demographics, past history, diagnosis, and New York Heart Association (HYHA) classification, and the operation mode, procedure duration, CPB duration, aortic clamp duration, intraoperative fluid balance, calculation of 24 hours postoperative fluid balance and vasoactive-inotropic score (VIS); 24 hours postoperative C-reactive protein (CRP), N-terminal B-type natriuretic peptide precursor (NT-proBNP) and PCT levels were collected. Two clinicians independently made the diagnosis of ARDS according to the Berlin definition, and the diagnosis was established only in patients with a consistent diagnosis. The differences in each parameter were compared between patients with moderate to severe ARDS and those without or with mild ARDS. Analysis of the ability of PCT to predict moderate to severe ARDS was evaluated by receiver operator characteristic curve (ROC curve). Multivariate Logistic regression was conducted to determine the risk factors of the development of moderate to severe ARDS. 108 patients were finally enrolled, including 37 patients with mild ARDS (34.3%), 35 patients with moderate ARDS (32.4%), 2 patients with severe ARDS (1.9%), and 34 patients without ARDS. Compared with patients with no or mild ARDS, patients with moderate to severe ARDS were older (years old: 58.5±11.1 vs. 52.8±14.8, P < 0.05), with a higher proportion of combined hypertension [45.9% (17/37) vs. 25.4% (18/71), P < 0.05], longer operative time (minutes: 363.2±120.6 vs. 313.5±97.6, P < 0.05), and higher mortality (8.1% vs. 0, P < 0.05), but there were no differences in the VIS score, incidence of acute renal failure (ARF), CPB duration, aortic clamp duration, and intraoperative bleeding, transfusion volume, and fluid balance between the two groups. Serum PCT and NT-proBNP levels in patients with moderate to severe ARDS at postoperative day 1 were significantly higher than those in patients with no or mild ARDS [PCT (μg/L): 16.33 (6.96, 32.56) vs. 2.21 (0.80, 5.76), NT-proBNP (ng/L): 2 405.0 (1 543.0, 6 456.5) vs. 1 680.0 (1 388.0, 4 667.0), both P < 0.05]. ROC curve analysis showed that the area under the curve (AUC) for PCT to predict the occurrence of moderate to severe ARDS was 0.827 [95% confidence interval (95%CI) was 0.739-0.915, P < 0.05]. When PCT cut-off value was 7.165 μg/L, the sensitivity was 75.7% and the specificity was 84.5%, for differentiating patients who developed moderate to severe ARDS from who did not. Multivariate Logistic regression showed that age and the elevated PCT concentration were independent risk factors for the development of moderate to severe ARDS [age: odds ratio (OR) = 1.105, 95%CI was 1.037-1.177, P = 0.002; PCT: OR = 48.286, 95%CI was 10.282-226.753, P < 0.001]. Patients with moderate to severe ARDS undergoing CPB cardiac surgery have a higher serum concentration of PCT than patients with no or mild ARDS. Serum PCT level may be a promising biomarker to predict the development of moderate to severe ARDS, the cut-off value is 7.165 μg/L.