Background and Purpose: Several methods have been proposed to predict prognosis in patients with paraquat poisoning, but all have their limitations. This retrospective study evaluated the usefulness of Acute Physiology and Chronic Health Evaluation (APACHE) II scores in risk stratification for patients with paraquat poisoning. Methods: Data from 58 adults with a diagnosis of paraquat poisoning presenting to a general hospital over a 10-year period were analyzed. APACHE II scores were calculated at 24 hours after admission and data on related parameters during the first 24 hours were collected for study. Results: The overall in-hospital mortality was 72.4% and mortality in the intensive care unit was 82.2%. APACHE II scores were higher in non-survivors (n = 42, 23.3 ± 12.4) than in survivors (n = 16, 6.7 ± 4.1; p ＜ 0.001). All 26 patients (44.8%) who received mechanical ventilation died. Of the 32 patients who received hemoperfusion, 25 (78.1%) died. Plasma paraquat concentration, estimated ingested amount of paraquat, and APACHE II score were significantly higher in non-survivors than in survivors (p ＜ 0.05 for all comparisons). There were significant correlations between APACHE II score and the following variables: plasma paraquat concentration, estimated ingested amount of paraquat, and the peak values during the first 24 hours after admission for fraction of inspired oxygen, alveolar-arterial oxygen gradient, aspartate aminotransferase, alanine aminotransferase, blood urea nitrogen, and serum creatinine (p ＜ 0.05 for all comparisons). Higher APACHE II score was associated with greater mortality. All patients who had an APACHE II score greater than 20 died before discharge. Conclusions: This study has demonstrated that the APACHE II score is positively correlated with plasma paraquat concentration and ingested amount of paraquat. An APACHE II score of 20 or higher is a good predictor of inhospital mortality.