BACKGROUND Acute pancreatitis in pregnancy (APIP) is a rare and serious condition, and severe APIP (SAPIP) can lead to pancreatic necrosis, abscess, multiple organ dysfunction, and other adverse maternal and infant outcomes. Therefore, early identification or prediction of SAPIP is important. AIM To assess factors for early identification or prediction of SAPIP. METHODS The clinical data of patients with APIP were retrospectively analyzed. Patients were classified with mild acute pancreatitis or severe acute pancreatitis, and the clinical characteristics and laboratory biochemical indexes were compared between the two groups. Logical regression and receiver operating characteristic curve analyses were performed to assess the efficacy of the factors for identification or prediction of SAPIP. RESULTS A total of 45 APIP patients were enrolled. Compared with the mild acute pancreatitis group, the severe acute pancreatitis group had significantly increased (P < 0.01) heart rate (HR), hemoglobin, neutrophil ratio (NEUT%), and neutrophil–lymphocyte ratio (NLR), while lymphocytes were significantly decreased (P < 0.01). Logical regression analysis showed that HR, NEUT%, NLR, and lymphocyte count differed significantly (P < 0.01) between the groups. These may be factors for early identification or prediction of SAPIP. The area under the curve of HR, NEUT%, NLR, and lymphocyte count in the receiver operating characteristic curve analysis was 0.748, 0.732, 0.821, and 0.774, respectively. The combined analysis showed that the area under the curve, sensitivity, and specificity were 0.869, 90.5%, and 70.8%, respectively. CONCLUSION HR, NEUT%, NLR, and lymphocyte count can be used for early identification or prediction of SAPIP, and the combination of the four factors is expected to improve identification or prediction of SAPIP.