Background: School based nutrition education programmes play a critical role in promoting positive dietary change in children. Psychosocial mediators of behaviour change such as behaviour intentions, nutrition knowledge and self-efficacy motivate children to change nutrition behaviour. Aim: To develop, implement and evaluate a school based nutrition education programme (NEP) tailored to the needs of seven to nine year olds living in resource limited settings in Pretoria, Gauteng Province, South Africa. Study design: Quasi experimental without a control group. Study setting: Two resource limited primary schools in Pretoria. Study participants: Learners, aged seven to nine years, in Grades 1-3 from the two selected schools that met the inclusion criteria. Informed consent and assent were obtained. Method: The NEP was developed based on analysis of the learners’ dietary practices data as reported by the parents and learners. These data were collected as part of the situational analysis for a larger study and was availed to the researcher on completion of the data collection. The qualitative domain was employed to determine the nutrition education (NE) needs of the learners. The study then proceeded in two phases; Phase 1, which resulted in the development of the NEP and NE materials as informed by the literature review and the outcome of the situational analysis and guided by the Social Cognitive Theory, the South African Food Based Dietary Guidelines and the six steps for designing a NEP by Contento. In Phase 2, the NEP was implemented with a conveniently selected sample of Grade 1, 2 and 3 learners from the two schools (School 1 and 2). All participants received nine nutrition lessons over six weeks. The quantitative domain was employed to evaluate the effects of the NEP on the food choice intentions of Grade 1 learners, and self-efficacy and nutrition knowledge of Grade 2 and 3 learners. Outcomes were measured at baseline, at six weeks and at 12 months respectively. A modified validated Pathways knowledge, attitudes and behaviour questionnaire was used. Paired t-test evaluated the effect of the NEP on the three outcomes and the independent samples test compared the differences between the schools and gender. The ANCOVA assessed the effect of school and gender on the measured outcomes, with the pre-assessment score as the covariate. The McNemar test was used to compare differences between related percentages. The repeated measures ANOVA was used to assess the effect of time on the measured outcomes. The statistical package SPSS version 22 was used to analyse the data and a significance level of 0.05 was employed. Ethical approval was sought from the Faculties of Education and Health Sciences of the University of Pretoria (Number: D2015/ 375A). Approval was also sought from the Gauteng Department of Basic Education (DoBE), as well as the primary schools that were involved in the study. Results: Situational analysis: The situational analysis revealed that the learners had unhealthy eating habits characterised by skipping breakfast, monotonous diets and high consumption of unhealthy energy dense foods in both the school and home environments. Some food groups such as legumes were completely missing from the learners’ diets. Evaluation of the NEP: Grade 1 (n=49) and Grade 2 and 3 learners (n=108), aged seven to nine years completed the study at six weeks and at 12 months. At post-assessment (six weeks), a significant improvement in the overall food choice intentions of Grade 1 learners was observed (M=0.41974 vs. M=0.5671; P<0.0001) (M=mean). Significant improvements were also observed in School 1 (P=0.001) and in School 2 (P=0.014) with a greater improvement in School 1. The girls in School 1 had significantly higher improvement in mean scores as compared to girls in School 2 (P=0.0001), while the boys in School 1 also had higher improvement in mean scores as compared to the boys in School 2, though not significant (P=0.275). However, at 12 months a significant decrease in the overall food choice intentions was observed from baseline to 12 months (M=0.436 vs. M=0.561 vs. M=0.446; P=0.0002) for Grade 1 learners. At six weeks, a decrease for overall self-efficacy mean scores (P=0.483) of Grade 2 and 3 learners were observed in School 1 (P=0.634) and School 2 (P=0.082), although not significant. School 1 had non significant higher mean self-efficacy scores as compared to School 2 (P=0.903). The improvement in the mean self-efficacy score was significantly higher for girls in School 1 as compared to the girls in School 2 (P=0.036). The boys in School 2 had higher non significant mean improvement of scores as compared to the boys in School 1 (P=0.351). At 12 months a significant decrease in overall self-efficacy mean scores was observed from baseline to 12 months (M=0.801 vs. M=0.791 vs. M=0.735; P=0.000) for the Grade 2 and 3 group as a whole.