Background: Theoretically, self-regulatory strategies (eg, goal setting, self-monitoring) are an important influence of behavior change, but very little research has examined the relationship between self-regulation and physical activity (PA) behavior. Petosa's (1993) 43-item PA self-regulation scale (PASR-43) affords the opportunity for studying this construct in the context of PA; however the PASR-43 has not been tested for structural aspects of validity. Therefore, this study examines the structural validity of the PASR-43 in older adults. Methods: The structural validity of the PASR-43 was tested in a large sample of older adults from North and South Carolina and Ohio (N = 460) using maximum likelihood estimation and confirmatory factor analysis in AMOS 5.0. Results: The original 6-factor model for the PASR-43 scale did not represent an acceptable fit to the data (x2 = 4732.25, df = 845, P < .0001, RMSEA = 0.10, NNFI = 0.67, CFI = 0.71). Based on a post hoc specification search, iterative model modifications resulted in a 12-item PA self-regulation scale (PASR-12) that represented an excellent fit to the data (x2 = 70.75, df = 39, P = .001, RMSEA = 0.04, NNFI = 0.98, CFI = 0.99). Conclusions: The PASR-12 provides a concise and valid measure of PA self-regulation for use with older adults. Future studies should cross-validate the PASR-12 and examine invariance across time and between age, ethnic, gender, and geo-graphical groups. © 2009 Human Kinetics, Inc.