To determine the optimal cut-off value of serum cotinine for distinguishing smoking status among Chinese adults based on a large-scale national sample. A cross-sectional study was conducted among 8 987 Chinese adults aged 20-79 years from 152 administrative counties across 31 provinces during 2017-2018. Sociodemographic characteristics, lifestyle, smoking status, and health status were collected via questionnaires and physical examinations. Blood samples were analyzed for serum cotinine levels using liquid chromatography-mass spectrometry and for blood creatinine levels using the picric acid method. Receiver operating characteristic (ROC) curve analysis was performed with serum cotinine concentration as the test variable and self-reported smoking status as the state variable. The optimal cut-off value was determined based on the maximum Youden's index, and the bootstrap method was used for repeated sampling (2 000 times) to evaluate the confidence interval of the cut-off value. The net reclassification index (NRI) was used to evaluate the discrimination ability of the cut-off value of this study, the cut-off value of the American population 1 (total population: 3.3 μg/L, men: 4.1 μg/L, women: 3.0 μg/L) and the cut-off value of the American population 2 (the recommended value of the United States Centers for Disease Control and Prevention for the total population: 10.0 μg/L) against the smoking status of the Chinese population. Statistical analyses were conducted using IBM SPSS 27 and Python 3.11, with a significance level of α=0.05. The age of the research subjects was (49.2±15.2) years. Among them, males accounted for 49.8% (4 477); smokers accounted for 28.8% (2 586); the detection rate of serum cotinine was 94.6% (8 501), and the () concentration of serum cotinine was 0.9 (0.3, 85.4) μg/L. The ROC curve analysis results showed that the cut-off value (95%) of serum cotinine in the total population was 8.8 (6.7-11.7) μg/L, with the specificity (95%) about 93.6%(92.7%-94.3%), the sensitivity (95%) about 91.0%(89.7%-92.3%) and the area under the curve (AUC) (95%) about 0.93 (0.92-0.94). The cut-off value (95%) of cotinine for males was 17.1 (8.8-21.8) μg/L, with the specificity (95%) about 90.7%(87.9%-92.0%), the sensitivity (95%) about 89.4%(88.4%-92.2%) and the AUC (95%) about 0.92 (0.91-0.93). The cut-off value (95%) of cotinine for females was 7.4 (3.3-15.0) μg/L, with the specificity (95%) about 95.6%(92.7%-96.8%), the sensitivity (95%) about 87.6%(81.6%-92.8%) and the AUC (95%) about 0.92 (0.87-0.95). The NRI analysis results showed that compared with the cut-off value of the American population 2, the NRI of this study's cut-off values in the total population, males and females were 0.020 (=0.015), 0.033 (=0.015) and 0.011 (=0.380), respectively, indicating that this study's cutoff value could have better classification performance in the total population and males. Compared with the cut-off value of the American population 2, the NRI of the total population in this study was 0.001 (=0.285). The serum cotinine cut-off value based on the analysis of large sample data in China is more suitable for distinguishing the smoking status of Chinese adults.