Eight tobacco samples were tested in cooperation centre for scientific research relative to tobacco（CORESTA）collaborative study by ten laboratories. The total alkaloids were extracted by water and acetic acid and determined by NM method and CRM35 method, respectively. The total alkaloids are ranging from 1% to 5%. The repeatability and reproducibility were calculated and comparative study was achieved by paired–t test, Bland–Altman analysis and Passing–Bablok regression. Results show that agreement results were obtained by NM and CRM35 whether water or acetic acid is used. The results with water extraction were higher than that with acetic acid extraction for both NM and CRM35, and the relative deviation was lower than 4%. Besides, there is a linear relationship between the results of water extraction and acetic acid extraction and the differences caused by extractions can be modeled by Passing–Bablok regression analysis. NM method avoid the use of KCN, and the results obtained were agreeable to those obtained by CRM35 method which used KCN, indicating NM is suitable for the determination of total alkaloid in tobacco.