Abstract: Postoperative pancreatic fistula (POPF) represents the most prevalent complication following pancreaticoduodenectomy, significantly impacting patient prognosis. In recent years, advances in quantitative CT parameter analysis, maturation of radiomics methodologies, and the widespread application of artificial intelligence (AI) have positioned noninvasive preoperative CT-based prediction of POPF risk as a prominent research focus. This article comprehensively reviews the latest research progress in CT parameters, radiomics, and AI for preoperative prediction of POPF, offering valuable insights for clinical practice and future research directions in the field.