Background: Different tests are available for diagnosing Helicobacter pylori infection. Aim: To compare the most commonly used tests either alone or in combination in Chinese patients with respect to routine clinical use or research purpose. Methods: A total of 294 consecutive dyspeptic patients without previous H. pylori treatment were recruited. During upper endoscopy, biopsies were taken from the antrum and corpus, for a commercially available CLO‐test, an in‐house rapid urease test, culture, polymerase chain reaction and histological examination. Patients then received a 13C‐urea breath test. The H. pylori status of each patient was determined by a concordance of test results. Results: For routine clinical use, histology (antral plus corpus biopsies) had an accuracy of 100%, whilst the rapid urease test had an accuracy of 99.7%. The 13C‐urea breath test was equally reliable, with an accuracy of 94.5%. Combinations of two tests did not confer additional advantage over the most accurate single test. For research purposes, the accuracy of using the criteria of two positives out of three diagnostic tests was 100% and equivocal results were not found. Conclusion: Histology with or without a rapid urease test was highly accurate for routine clinical use. Alternatively, the 13C‐urea breath test was an equally reliable non‐invasive test. The two positives out of three tests approach was highly reliable in predicting H. pylori status of untreated Chinese patients in a research setting.