This study tested the effects of two theory-based interventions to increase fruit and vegetable intake. Hypothesized intervention mediators included self-efficacy (SE), social support (SS), autonomous motivation (AM), and controlled motivation (CM). At baseline, 1,021 African American adults were recruited from 16 churches randomized to one comparison and two intervention groups: Group 1 (standard educational materials), Group 2 (culturally targeted materials), and Group 3 (culturally targeted materials and telephone-based motivational interviewing). A well-fitted model based on structural equation modeling--chi[superscript 2](df = 541, N = 353, 325) = 864.28, p less than 0.001, normed fit index = 0.96, nonnormed fit index = 0.98, comparative fit index = 0.98, root mean square error of approximation = 0.042--demonstrated that AM was both a significant mediator and moderator. In the subgroup with low baseline AM, AM mediated 17% of the effect of the Group 3 intervention on fruit and vegetable intake. Conversely, SS, SE, and CM were not significant mediators. Implications related to theory and intervention development are discussed. (Contains 2 tables and 2 figures.)