Background: Improving children's fruit and vegetable (FV) preferences may be important as preferences can predict FV consumption. The purpose of this study was to evaluate FV preferences over time, with repeated experience, as part of the "Fresh Fruit and Vegetable Program" (FFVP). Methods: Fruits (F; N = 28) and vegetables (V; N = 29) were distributed twice a week, over 35 weeks, at a participating FFVP school (N = 236 students, 12 teachers, K-2nd grade). Preference ratings using 3-point Likert scale were analyzed over 35 weeks. Results: For 57 FVs rated for preference, ratings revealed that F had higher frequency of children choosing "I like it" than for V (78% F; 38.2% V; p < 0.05) and liking distribution was different between F and V (p < 0.001). Significant relations were found between liking and: (1) grade (r = -0.02, p = 0.02), and (2) time (r = -0.09, p < 0.001). Models indicated that V served ([beta] = -0.40), timepoint ([beta] = -0.07), and grade level ([beta] = -0.02) accounted for significant variance for preference ratings (R[superscript 2] = 0.17, p < 0.001), indicating that preference ratings declined over time. Conclusions: Fruits were preferred over vegetables. Overall preference ratings were negatively impacted by time, grade level, and vegetables served. Being exposed one time to a variety of FVs did not improve ratings for vegetables.