The transtheoretical model (TTM) was developed as a guide for understanding behavior change. Little attention has been given, however, to the appropriateness of the TTM for explaining the adoption of exercise behavior in older adults. The purposes of this study were to determine the reliability of the TTM instruments and validate TTM predictions in 86 community-dwelling older adults (mean age 75.1 +/- 7.0 years, 87% women) who were participants in a 16-week walking program. TTM construct scales--self-efficacy, decisional balance (pros and cons), and processes of change (behavioral and cognitive)--were generally reliable (all>.78). Behavioral processes of change increased from baseline to follow-up, but pros, cons, and cognitive processes did not change among participants who became regular exercisers. Stage of change did not predict exercise adoption, but baseline self-efficacy predicted walking behavior. These results lend partial support to the TTM in predicting exercise behavior.