Due to the lack of span test for the use in language-specific and cross-language studies, this study provides L1 and L2 researchers with a reliable language-independent span test (math span test) for the measurement of working memory capacity. It also describes the development, validation, and scoring method of this test. This test included 70 simple math problems, and was developed based on Salthouse and Babcock's (1991) and Robert and Gibson's (2002) math span tests. The shortcomings of the test were identified and removed over five pilot studies on 48 participants. The final test was used in an experimental study with a group of L1 Persian EFL learners. Results of an item analysis, as indicated by Cronbach's Alpha, indicated an internal reliability of 0.850 and 0.863 for the math span test processing and recall respectively. This suggests that the newly developed test is reliable enough and could be used to measure working memory capacity in L1 and L2 studies. This study also provides a clear procedure for the development and scoring of a math span test for the use in L1 and L2 studies.