Several methods have been used for the detection of Helicobacter pylori (HP) infection. However, few reports have evaluated the accuracy of each method and compared the level of HP infection. HP infection was evaluated semi-quantitatively using immunohistochemical staining and accuracy of several methods to detect HP infection were compared. Biopsy specimens, obtained from a total 64 patients who underwent endoscopy for evaluation of gastroduodenal diseases, were studied using a rapid urease test, culture method, and immunohistochemical method. The infection was graded according to the number of the individual bacteria seen in a highly magnified visual field, and defined as follows: (0) = 0; (1+) < 10; (2+) = 10-29; (3+) = 30-99, (4+) > 100. The rapid urease test had a sensitivity of 53%, specificity of 100%, and accuracy of 73%. The culture method had a sensitivity of 75%, specificity of 100%, and accuracy of 86%. Sensitivity of rapid urease test and the culture method decreased in a positive correlation to the decrease in total number of HP bacteria counted. Using the rapid urease test, sensitivity was < 30% when the grade of HP infection was (2+) or less, while 100% sensitivity was obtained only when the grade of HP infection was (4+). On the other hand, sensitivity of culture method, remained between 80 and 90% when HP infection was (2+) or more. The rapid urease test and culture of HP may result in false negative tests for mild infection although those have high sensitivity and specificity for moderate to severe infection. Immunohistochemical stain provides a reliable semi-quantitative diagnosis of HP infection. Clinicians should be aware of the characteristics of each method to detect HP infection, and select the appropriate ones for their purpose.