Albeit their practical implications, psychosocial correlates of physical activity and related weight loss in the treatment of extreme obesity have been only sparsely addressed in adults; and even more minimally focused upon in adolescents. This research contrasted results of a 6-month social cognitive theory-based physical activity-support protocol along with standard nutrition counseling in groups of adolescents (n = 19; agemean = 15.4 years) and adults (n = 26, agemean = 44.7 years) with class 3 (extreme) obesity [overall body mass index (BMI)mean = 53.4 kg/m2, standard deviation (SD) = 8.2]. Although baseline total mood disturbance scores were significantly greater in both groups when contrasted with age-corresponding normative values, between-group scores did not significantly differ. BMI and physical activity changes over 3 and 6 months were significantly more favorable in the adult group. There were significant overall improvements in exercise self-efficacy, physical activity self-regulation and total mood disturbance scores, however, the self-efficacy and self-regulation increases were significant in only the adult group. Inverse relationships between changes in physical activity and BMI were stronger in the adult group. Regardless of group, simultaneous entry of changes in exercise self-efficacy, physical activity self-regulation and total mood disturbance scores, over both 3 and 6 months, significantly predicted physical activity changes (R2-values = 0.45–0.75, p-values < 0.001). In each model, self-regulation change was the most salient of those three psychosocial predictors. Although findings suggested that psychosocial correlates of physical activity, leading to weight change, are similar in adolescents and adults with extreme obesity, facilitation of larger effect sizes and/or foci on additional theory-based determinants may be required for clinically meaningful treatment outcomes in adolescents.