The purpose of this study was to determine the effects of training change-of-direction speed and small-sided games on performance in the Planned-AFL agility test and reactive agility. Twenty-five elite-standard U-18 Australian Rules football players were randomly allocated either to a change-of-direction group or a small-sided games group. Players participated in one or two 15-min sessions per week with 11 sessions conducted over a 7-week period during the season. Tests conducted immediately before and after the training period included the Planned-AFL agility test and a video-based reactive agility test specific to Australian Rules football. The reactive agility test variables were total time, decision time and movement response time. The small-sided games group improved total time (P = 0.008, effect size = 0.93), which was entirely attributable to a very large reduction in decision time (P < 0.001, effect size = 2.32). Small-sided games produced a trivial change in movement response time as well as in the Planned-AFL agility test (P > 0.05). The change-of-direction training produced small to trivial changes in all of the test variables (P > 0.05, effect size = 0-0.2). The results suggest that small-sided games improve agility performance by enhancing the speed of decision-making rather than movement speed. The change-of-direction training was not effective for developing either change-of-direction speed as measured by the Planned-AFL test or reactive agility.