This paper assesses the impact of school feeding programs on internal effectiveness (dropout and grade repetition) of public primary schools in Senegalese rural area using a “randomized trial”. Our study sample consists of 120 schools, which did not have school canteens before the intervention. These schools were selected in four poor regions of Senegal from 2009 to 2010. They were assigned randomly (60 schools in each group) to control and treatment groups. Students in second and fourth grade were observed in each of the selected schools. Different tests (Levene, Student, Kolmogorov-Smirnov and Mann-Whitney) are used to verify whether the two groups are similar or no before the beginning of the intervention. The results show that there are no difference on observed characteristics between the treated and the control groups. We use the difference-in-difference method to evaluate the impact of the intervention on student's dropout and grade repetition. The results of the estimation reveal that school feeding does not reduce significantly grade repetition. In contrast, we find significant decrease in school dropout, particularly for girls. Also, providing foods to children at school reduces early dropout for all students with a larger impact on girls.