Researchers interested in the relationship between working memory and interpreting tend to adopt working memory (WM) span tasks developed by psychologists as WM measurement tools. Among the most commonly used are those that consist of words and sentences such as the reading span task. The reliability of such tasks has been verified in several studies in cognitive psychology. However, no study has been found to demonstrate that reliability is maintained when the WM span tasks are administered to interpreters. This study aimed to examine and compare the reliability of reading span tasks which involve linguistic elements and that of an operation span task, a language-independent WM span task. For this purpose, student interpreters were invited to perform the WM span tasks and then the equivalent form test reliability and the test-retest reliability of the tasks were measured. The results suggest that only the operation span task is a reliable tool for testing the WM capacity of interpreters.