To establish a modified controlled abciximab and device investigation to lower late angioplasty complication (CADILLAC) score, and to compare the predictive value of modified CADILLAC score, the global registry of acute coronary event (GRACE) score and the thrombolysis in myocardial infarction (TIMI) score in predicting the risk of short-term death after percutaneous coronary intervention (PCI) in patients with acute ST segment elevation myocardial infarction (STEMI). A retrospective study was conducted. The clinical data of 169 STEMI patients under going PCI admitted to the department of cardiology of Guizhou Provincial People's Hospital from September 2019 to December 2020 through emergency chest pain fast track were enrolled. A multivariate Logistic regression analysis was used to screen the factors closely related to the mortality risk within 30 days of STEMI, and a modified CADILLAC scoring system was established by referring to CADILLAC scoring settings. The score of modified CADILLAC, GRACE and TIMI scores of patients were calculated after admission, and the number of deaths due to cardiovascular disease (CVD) within 30 days after onset was recorded. The receiver operating characteristic curve (ROC curve) was used to evaluate the predictive value of three scoring systems on the risk of death within 30 days after PCI in patients with STEMI. In 169 STEMI patients, 16 patients died of CVD within 30 days after PCI, and the actual case mortality was 9.47%. Multivariate Logistic regression analysis showed that age > 75 years old, cardiac function Killip ≥ Grade III, ventricular arrhythmia, ST segment elevation ≥ 0.2 mV, cardiac troponin I (cTnI) increase, systolic blood pressure (SBP) < 90 mmHg (1 mmHg ≈ 0.133 kPa) were all independent predictors of death after PCI in STEMI patients. The improved CADILLAC scoring system was constructed based on the above predictive factors combined with left ventricular ejection fraction (LVEF) less than 0.40. The GRACE, TIMI and modified CADILLAC scores of dead patients were significantly higher than those of survival patients (GRACE score: 197.60±31.83 vs. 149.81±36.72, TIMI score: 11.21±2.13 vs. 7.27±1.97, modified CADILLAC score: 12.60±2.52 vs. 6.96±2.17, all P < 0.05). The higher the risk stratification of the three scores, the higher the mortality of patients with CVD within 30 days after PCI [the mortality of patients with low, medium and high risk in GRACE score were 2.41% (2/83), 9.61% (5/52) and 26.47% (9/34); the mortality of patients with low, medium and high risk in TIMI score were 3.12% (3/96), 12.82% (5/39) and 23.53% (8/34); and the mortality of patients with low, medium and high risk in modified CADILLAC score were 3.19% (3/94), 7.69% (4/52) and 39.13% (9/23), respectively, all P < 0.01]. The area under the ROC curve (AUC) of the GRACE, TIMI and the modified CADILLAC scores predicting the risk of death 30 days after PCI in STEMI patients were 0.855 [95% confidence interval (95%CI) was 0.702-0.923], 0.725 (95%CI was 0.666-0.812) and 0.882 (95%CI was 0.732-0.936), respectively, all P = 0.000; the sensitivity of its prediction accuracy were 81.59%, 78.65% and 89.26%, and the specificity were 78.62%, 57.12% and 75.54%, respectively. The GRACE and the modified CADILLAC scores have predictive value for the short-term mortality risk of STEMI patients after PCI, and the modified CADILLAC score is more accurate. But the TIMI score has a poor predictive effect on the short-term mortality risk of STEMI patients after PCI.