BACKGROUND: Age at menarche is the reflection of cumulative pre-adolescent exposure of girls to either adverse environment such as food insecurity or affluent living conditions. Food insecurity could result in inadequate nutrient intake and stress, both of which are hypothesized to have opposing effects on the timing of menarche through divergent pathways. It is not known whether food insecure girls have delayed menarche or early menarche compared with their food secure peers. In this study we test the competing hypothesis of the relationship between food insecurity and age at menarche among adolescent girls in the Southwest Ethiopia. METHODS: We report on 900 girls who were investigated in the first two rounds of the five year longitudinal survey. The semi-parametric frailty model was fitted to determine the effect of adolescent food insecurity on time to menarche after adjusting for socio-demographic and economic variables. RESULTS: Food insecure girls have menarche one year later than their food secure peer (median age of 15 years vs 14 years). The hazard of menarche showed a significant decline (P = 0.019) as severity of food insecurity level increased, the hazard ratio (HR) for mild food insecurity and moderate/severe food insecurity were 0.936 and 0.496, respectively compared to food secure girls. Stunted girls had menarche nearly one year later than their non-stunted peers (HR = 0.551, P < 0.001). CONCLUSION: Food insecurity is associated with delay of age at menarche by one year among girls in the study area. Stunted girls had menarche one year later than their non-stunted peers. Age at menarche reflects the development of girls including the timing of sexual maturation, nutritional status and trajectory of growth during the pre-pubertal periods. The findings reflect the consequence of chronic food insecurity on the development and well-being of girls in the study area.