Background: Compared with other types of acute pancreatitis (AP), hypertriglyceridemic acute pancreatitis (HTG-AP) is younger, recurrent and more prone to exacerbation. Severe HTG-AP has a high fatality rate. Early and accurate prediction of the severity is crucial. However, there is currently a lack of a specific scoring system for the severity of HTG-AP. Aim/Purpose: To construct a risk prediction model that can accurately predict severe HTG-AP in the early stage and evaluate its clinical value. Methods: The clinical data of 1768 patients with AP admitted to the Second Affiliated Hospital of Anhui Medical University from January 2020 to May 2023 were analyzed retrospectively, and 136 HTG-AP patients were finally selected. Univariate and multivariate analysis were performed for the early onset indicators to identify the independent risk factors for developing SAP in the patients of HTG-AP. Logistic regression was then utilized to establish a risk prediction model for the severity of HTG-AP, which was subsequently evaluated for its performance through discrimination and calibration analysis. Results: Of the 136 patients with HTG-AP, 39 patients (28.7%) progressed to severe acute pancreatitis (SAP). Multivariate analysis revealed that CRP, RDW/SC, and D-dimer were independent risk factors for developing SAP in the patients of HTG-AP. The logistic regression analysis to establish prediction model was: Logit P = − 8.101 + 0.008 × CRP + 0.425 × D-dimer + 0.743 × RDW/SC. The receiver-operating characteristics (ROC) curve showed that area under curve (AUC) value of CRP, RDW/SC, D-dimer, and the prediction model were 0.831, 0.843, 0.874, and 0.915, respectively. Moreover, the AUC value of the prediction model and commonly used scoring systems of AP were compared: prediction model (AUC = 0.915) > Ranson (AUC = 0.900) > SOFA (AUC = 0.899) > CTSI (AUC = 0.889) > BISAP (AUC = 0.887). Conclusion: CRP, RDW/SC and D-dimer were independent risk factors for SAP in the patients of HTG-AP. Compared with commonly used scoring systems of AP, the prediction model had good clinical prediction ability, providing reference for early identification of the patients developing severe HTG-AP and active intervention. [ABSTRACT FROM AUTHOR]