The reliability and validity of a soccer-specific field test of repeated sprint ability was assessed. Seven male games players performed the repeated sprint test on six separate occasions. The temporal pattern of the mean sprint time was analysed by using coefficient of variation with confidence intervals (CI), and repeated measures ANOVA. A within subject mean coefficient of variation of 1.8% (95% CI, 1.5-2.4) was found for performance in the repeated sprint test across all six trials. The mean coefficient of variation across trials 2-4 was found to be 1.9% (95% CI, 1.3-3.1), compared to trials 4-6, where it was 1.4% (95% CI, 1.0-2.3). The ANOVA showed that a significant difference was present between the trials (F6,30 9.8. P<0.001). A Tukey post-hoc test showed that significant differences were present between trial I and trials 3-6, and trial 2 and trial 5. The learning effect was complete by trial 3. Performance in the repeated sprint test was compared to total running time averaged from two repeats of the maximal anaerobic running test laboratory protocol. Mean sprint time in the repeated sprint test and total running time in the laboratory protocol had a correlation coefficient of r = -0.298 (P = 0.516, n = 7), suggesting that the energetics of the two tests are not closely related. In conclusion, this soccer-specific field test demonstrated high reliability.