To study the optimal pain control goal for preventing delirium in critical patients. A prospective cohort study were conducted. The patients admitted to general departments and transferred to the intensive care unit (ICU) due to critical illness in the First People's Hospital of Changde from January 2017 to November 2019 were enrolled. The General data of the patients were collected within 48 hours after admission. All patients admitted to the ICU were evaluated for pain level using the critical care pain observation tool (CPOT) every 8 hours by nurses, and confusion assessment method of ICU (CAM-ICU) was used to screen delirium patient every 8 hours by the leader of nursing team without knowing the pain level of the patients, until the subjects were transferred out of ICU. The receiver operating characteristic (ROC) curve was drawn, the area under ROC curve (AUC) and the optimal threshold were analyzed with delirium as the reference standard; according to the optimal threshold, multivariate Logistic regression analysis was used to evaluate the correlation between CPOT score and delirium. During the study period, 575 patients were admitted to the participating departments and passed the preliminary screening according to the inclusion and exclusion criteria. During the study period, 34 patients were excluded due to incomplete data. Finally, a total of 541 patients were enrolled in the analysis, including 149 patients in delirium group and 392 patients in non-delirium group. There was no significant difference in gender, age, source of patients, education level, smoking history, drinking history, family mental history, acute physiology and chronic health evaluation II (APACHE II) score or other general information between the two groups. There were 10.1% (15/149) of patients in the delirium group used opioids, which was significantly higher than 4.3% (17/392) in the non-delirium group, and the difference was statistically significant (P < 0.05). The CPOT score in the delirium group was significantly higher than that in the non-delirium group (4.24±1.78 vs. 2.75±1.95, P < 0.01). The patients were subdivided into young group (< 40 years old), middle-aged group (40-65 years old) and old group (> 65 years old) according to age. The analysis results were consistent with the overall analysis results. ROC curve analysis showed that the AUC of CPOT score predicting delirium was 0.719; when the best threshold value of CPOT score was 2.5, the sensitivity was 91.3%, the specificity was 49.0%, the positive predictive value was 40.5% and the negative predictive value was 93.7%. Multivariate Logistic regression analysis showed that the risk of delirium in ICU patients with CPOT score ≥ 3 was 10.043 times higher than that in patients with CPOT score < 3 [odds ratio (OR) = 10.043, 95% confidence interval (95%CI) was 5.498-18.345, P < 0.001]. When the gender, age, APACHE II score, smoking history, drinking history, opioids usage were adjusted, the risk of delirium in patients with CPOT score ≥ 3 was 10.719 times higher than that in patients with CPOT score < 3 (OR = 10.719, 95%CI was 5.689-20.196, P < 0.001). The best pain control goal for preventing the occurrence of delirium in ICU patients is a CPOT score of 3 or less.