The purpose of the present study was to examine the effects of varying exercise intensities and changes in self-efficacy on anxiety reduction in a sample of healthy, older adults. Eighty older adults from a randomized controlled exercise trial participated in this study and completed measures of self-efficacy and the State Anxiety Inventory (SAI) prior to and following light-, moderate-, and high-intensity exercise. Latent growth curve modeling analyses revealed that although anxiety was reduced following the light-intensity condition, no significant changes in anxiety occurred following the moderate-intensity condition, and anxiety increased following the high-intensity condition. In addition, changes in self-efficacy were related to anxiety responses only in the moderate-intensity condition. An analysis of SAI items indicated that although the Eight-intensity condition resulted in decreased arousal and anxiousness, the high-intensity condition resulted in increased arousal and decreased anxiousness. These results are discussed in terms of social cognitive theory and the appropriateness of the SAI for use in exercise settings.