The EAT–Lancet Commission promulgated a universal reference diet. Subsequently, researchers constructed an EAT–Lancet diet score (0–14 points), with minimum intake values for various dietary components set at 0 g/d, and reported inverse associations with risks of major health outcomes in a high-income population. We assessed associations between EAT–Lancet diet scores, without or with lower bound values, and the mean probability of micronutrient adequacy (MPA) among nutrition-insecure women of reproductive age (WRA) from low- and middle-income countries (LMIC). We analysed single 24-h diet recall data (n 1950) from studies in rural DRC, Ecuador, Kenya, Sri Lanka and Vietnam. Associations between EAT–Lancet diet scores and MPA were assessed by fitting linear mixed-effects models. Mean EAT–Lancet diet scores were 8·8 (SD 1·3) and 1·9 (SD 1·1) without or with minimum intake values, respectively. Pooled MPA was 0·58 (SD 0·22) and energy intake was 10·5 (SD 4·6) MJ/d. A one-point increase in the EAT–Lancet diet score, without minimum intake values, was associated with a 2·6 (SD 0·7) percentage points decrease in MPA (P < 0·001). In contrast, the EAT–Lancet diet score, with minimum intake values, was associated with a 2·4 (SD 1·3) percentage points increase in MPA (P = 0·07). Further analysis indicated positive associations between EAT–Lancet diet scores and MPA adjusted for energy intake (P < 0·05). Our findings indicate that the EAT–Lancet diet score requires minimum intake values for nutrient-dense dietary components to avoid positively scoring non-consumption of food groups and subsequently predicting lower MPA of diets, when applied to rural WRA in LMIC.