BACKGROUND: More than 700,000 ethnic Rohingya have crossed the border from Rakhine State, Myanmar to Cox's Bazar District, Bangladesh, following escalated violence by Myanmar security forces. The majority of these displaced Rohingya settled in informal sites on previously forested land, in areas without basic infrastructure or access to services. METHODS AND FINDINGS: Three cross-sectional population-representative cluster surveys were conducted, including all informal settlements of Rohingya refugees in the Ukhia and Teknaf Upazilas of Cox's Bazar District. The first survey was conducted during the acute phase of the humanitarian response (October-November 2017), and the second and third surveys were conducted 6 (April-May 2018) and 12 (October-November 2018) months later. Anthropometric indices (weight, height, mid-upper arm circumference [MUAC], oedema) and haemoglobin (Hb) were measured in children aged 6-59 months following standard procedures. Final samples for survey rounds 1, 2, and 3 (R1, R2, and R3) included 1,113, 628, and 683 children, respectively, of which approximately half were male (50.7%-53.5% per round) and a third were 6-23 months of age (32.4%-33.3% per round). Prevalence of global acute malnutrition (GAM) as assessed by weight for height in R2 (12.1%, 95% CI: 9.6-15.1) and R3 (11.0%, 95% CI: 8.4-14.2) represent a significant decline from the observed prevalence in R1 (19.4%, 95% CI: 16.8-22.3) (p < 0.001 for both comparisons). Overall, the prevalence of anaemia significantly declined (p < 0.001) between the first 2 rounds (47.9%, 95% CI: 44.1-51.7 and 32.3%, 95% CI: 27.8-37.1, respectively); prevalence increased significantly (p = 0.04) to 39.8% (95% CI, 34.1-45.4) during R3 but remained below R1 levels. Reported receipt of both fortified blended foods (12.8%) and micronutrient powders (10.3%) were low during R1 but increased significantly (p < 0.001 for both) within the first 6 months to 49.8% and 29.9%, respectively. Although findings demonstrate improvement in anthropometric indicators during a period in which nutrition programme coverage increased, causation cannot be determined from the cross-sectional design. CONCLUSIONS: These data document significant improvements in both acute and micronutrient malnutrition among Rohingya children in makeshift settlements. These declines coincide with a scaleup of services aimed at prevention and treatment of malnutrition. Ongoing activities to improve access to nutritional services may facilitate further reductions in malnutrition levels to sustained below-crisis levels.