Childhood obesity is increasing in the United States; thus, physicians, nurses, and other health care professionals seek to refer patients to interventions that will reliably improve physical activity and nutrition behaviors. The present 12-week, two-session-per-week protocol, based on social cognitive theory, was given preliminary testing with 23 obese children (Mage = 11.7 years) with risk factors for Type 2 diabetes. A significant within-group improvement in number of days per week of 60 or more minutes of voluntary physical activity was reported. Changes in measures of both task self-efficacy (β = .39) and self-regulatory efficacy (β = .44) significantly contributed to the significant portion of the variance explained in change in voluntary physical activity (R2 = .40). Significant improvements in total cholesterol and body mass index (kg/m2) were also found. Correlations between changes in physical activity and changes in each physiological factor tested were each in the expected direction but did not reach statistical significance. Results suggest that replications and extensions of this pilot study, with greater experimental power, are warranted. © 2010 Elsevier Inc.