In Japan, both a stick-type kit and an enzyme-linked immunosorbent assay (ELISA) kit are available for the detection of antibodies to Helicobacter pylori in urine. However, the accuracy of these tests has not been fully examined in northern Japanese populations. Urine samples from 359 subjects were tested using a stick-type H. pylori-antibody detection kit (RAPIRUN), and urine samples from 201 subjects were tested using an ELISA-based test (URINELISA). The prevalence of H. pylori infection was determined by the 13C-urea breath test (UBT) and a monoclonal antibody-based stool antigen test (TPAg). Subjects were considered to have the infection if either the UBT or rapid TPAg results were positive. The percentage of positive test results for RAPIRUN and URINELISA was 54.0% and 40.8%, respectively. Sensitivity and specificity were 83.3% and 67.0%, respectively, for RAPIRUN and 86.5% and 85.8% for URINELISA. Nineteen subjects had cut-off index values of between 0.4 and 0.9 by URINELISA, and 4 of these subjects (21.1%) were found to be infected with H. pylori. The urine-based ELISA was more accurate than the rapid stick-type kit in these patients. If negative ELISA results are near the cut-off value, subjects should receive an additional test to determine whether they are infected with H. pylori.