The purpose of this study was to test Hardy and Fazey's (1988) catastrophe model of anxiety and performance. Subjects were 20 male basketball varsity players recruited from two junior colleges. On each team, subjects were randomly assigned to eight high or low cognitive anxiety group. Prior to the experiment, subjects' were asked to complete CSAI-2 which assessed their state anxiety. Then each subject was asked to performed jump shots at the fault line. Levels of physiological arousal ranged from subject's predicted maximum heart rate (M, M=220-age) down to M-105. Heart rate bandwidths were constructed by increments of 15 bits per minute. In order to control for fatigue effects, the subjects were randomly divided into two groups. One group performed the criterion task with heart rate bandwidths increasing to M and then decreasing to M-105, while the other group performed the task with heart rate bandwidths decreasing from M down to M-105 and then increasing back up to M. In order to manipulate heart rate, subjects were asked to performed shuttle run or took rest until their heart rate meet the criterion. Subject's heart rate was monitored by using a Polar Electro Sport Tester. Subject's performance was measured by using a scoring system of 5 for a clean basket; 4 for a rim and in; 3 for backboard and in; 2 for a rim and out; 1 for backboard and out; and 0 for a complete miss. The results of this study partically supported the catastrophe model. It was found that a inverted-U relationship between physiological arousal and performance under low cognitive anxiety condition. Hysteresis was found only under high cognitive anxiety condition. In addition, subjects' best performance under high cognitive anxiety condition was significantly better than under low cognitive anxiety condition. However, there was no significantly differences between high cognitive anxiety condition and low cognitive anxiety condition in terms of subjects' worst performance. The implications and future directions of this study are discussed.