In this study we developed a novel badminton speed test (BST). The test was designed to mimic match play. The test starts in the center of the court and consists of five maximal actions to sensors located in each of the four corners of the court. The 20 actions are performed in randomized order as dictated by computer screen shots displayed one second following completion of the previous action. We assessed day-to-day variation in elite players and specificity of the test was evaluated by comparing 30 meter sprint performance and time to complete the BST in 20 elite, 21 skilled players and 20 age-matched physical active subjects (non-badminton players). Sprint performance was similar across groups, whereas the elite players were significantly (P<0.05) faster in the BST (total test time (32.3 ± 1.1) s; average 1.6 s per action) than the skilled (34.1 ± 2.0) s and non-badminton players (35.7 ± 1.7) s. Day-to-day coefficient of variation (CV) of the BST Test was 0.7 % for the elite players, while CV for repeated tests on the same day was 1.7 % for elite, 2.6 %, for skilled and 2.5 % for non-badminton players. On this basis we suggest that the BST may be valuable for evaluation of short-term maximal movement speed in badminton players. Thus, the BST appears to be sport specific as it may discriminate between groups (elite, less trained players and non-badminton players) with similar sprinting performance and the low test-retest variation may allow for using the BST to evaluate longitudinal changes e.g. training effects or seasonal variations.