Successful athletes are better at performing efficiently than the inferior in particular sports scenarios, while most existing performance tests in the field do not cover the sport-specific context fully. There were two purposes in this study: 1) to evaluate the reliability and validity of a novel Sector Reactive Agility Test (SRAT) which mimicked a reactive-agility defensive scenario in Touch, and 2) to determine the relationships between Touch players' agility and sprint performance. Twenty male Touch players from the elite division and another 20 from the amateur division were invited to participate in this study. They performed SRAT and a 20-m sprint test in two days. Excellent reliability and high precision were found in SRAT (intraclass correlation coefficient [ICC] = 0.97) and 20-m sprint test (ICC = 0.91). The time of completion in SRAT of the elite Touch players (23.93 s) was 2.95 s significantly shorter than that of the amateur players with a large effect size. Elite Touch players also demonstrated moderately faster (0.11 s) than the amateur Touch players in the 20-m sprint test. SRAT demonstrated high test-retest reliability and accuracy in measuring reactive-agility performance in Touch. The minimal detectable changes in SRAT and 20-m sprint test were 1.04 s and 0.13 s respectively. Furthermore, the speed of the 20-m sprint test and playing experience were associated with the time of completion of SRAT, explaining 56% of its variance (p < 0.001). Other factors, such as cognition and the ability to control own central gravity, are deemed possible to influence Touch players' agility. Therefore, SRAT should be adopted in Touch player selection and training monitoring.