Social-cognitive theory (SCT) was used to explain the fruit and vegetable intake of 1,398 3rd graders. SCT variables assessed included self-efficacy, outcome expectations, preferences, social norms, asking skills, and knowledge. Fruit and vegetable intake was assessed with 7-day records. Bivariate correlations with fruit and vegetable intake ranged from .17 for asking skills to .29 for fruit and vegetable preferences. In analyses controlling for school-level clustering, only preferences and positive outcome expectations remained significantly associated with fruit and vegetable intake, accounting for approximately 10%-11% of the variance. Limitations in the conceptualization, scope, and measurement of the variables assessed may have contributed to the weak associations observed. Models incorporating factors other than individual-level social-cognitive variables may be required to more fully explain children's dietary behavior.