Objectives. We assessed the impact of a rewards-based incentive program on fruit and vegetable purchases by low-income families. Methods. We conducted a 4-phase prospective cohort study with randomized intervention and wait-listed control groups in Philadelphia, Pennsylvania, in December 2010 through October 2011. The intervention provided a rebate of 50% of the dollar amount spent on fresh or frozen fruit and vegetables, reduced to 25% during a tapering phase, then eliminated. Primary outcome measures were number of servings of fruit and of vegetables purchased per week. Results. Households assigned to the intervention purchased an average of 8 (95% confidence interval [CI] = 1.5, 16.9) more servings of vegetables and 2.5 (95% CI = 0.3, 9.5) more servings of fruit per week than did control households. In longitudinal price-adjusted analyses, when the incentive was reduced and then discontinued, the amounts purchased were similar to baseline. Conclusions. Investigation of the financial costs and potential benefits of incentive programs to supermarkets, government agencies, and other stakeholders is needed to identify sustainable interventions.