The present study was to verify the difference of the test-retest reliability between criterion-referenced standard test and norm referenced standard test applied to the three different long distance running(800m, 1200m, and 1600m) tests by sex and grade. In order to accomplish this purpose, the study selected 208 male and 241 female students in grade 1, 2, and 3 from middle school as subjects. The trial order was based on counterbalanced design. The experimental design of the study was 2 × 3 × 2 × 3(sex × grades × trials × test methods) ANOVA repeated measure design. The study suggested that long-distance running test was reliable subtest for evaluating students` physical fitness. The results were as follows : 1. Long-distance running test was more suitable measure for the men than the women, and for the high grade than the low grade in terms of norm referenced standards but long-distance running test was not suitable measure for the both sex in terms of criterion-referenced standards, therefore this study was ahead of investigating more validity on long-distance running tests related to establishment of cutoff score. 2. 1600m run was the most reliable test for evaluating physical fitness among long-distance running test(800m, 1200m and 1600m) regardless sex and grade. In the case of grade 3,800m run was the highest coefficient among 800m, 1200m, and 1600m in the reliablity of norm-referenced standards but there was not statistically significant difference among 800m, 1200m and 1600m run. It was concluded that 1600m run was utilized as the most reliable test for evaluating; physical fitness in middle school.