\nObjective: The purpose of this paper is to examine the relationship between childhood undernutrition and poverty in urban and rural areas.\nDesign: Anthropometric and socio-economic data from Multiple Indicator Cluster Surveys in Angola-Secured Territory (Angola ST), Central African Republic and\nSenegal were used in this analysis. The population considered in this study is children 0 –59 months, whose records include complete anthropometric data on height, weight, age, gender, socio-economic level and urban or rural area of residence. In addition to simple urban/rural comparisons, the population was strati?ed using a wealth index based on living conditions and asset ownership to compare the prevalence, mean Z-score and odds ratios for stunting and wasting.\nResults: In all cases, when using a simple urban/rural comparison, the prevalence of stunting was signi?cantly higher in rural areas. However, when the urban and rural populations were strati?ed using a measure of wealth, the differences in prevalence of stunting and underweight in urban and rural areas of Angola ST, Central African Republic and Senegal disappeared. Poor children in these urban areas were just as likely to be stunted or underweight as poor children living in rural areas. The odds ratio of stunting in the poorest compared with the richest quintile was 3.4, 3.2 and 1.5 in Angola ST, Senegal and Central African Republic, respectively.\nConclusions: This paper demonstrates that simple urban/rural comparisons mask wide disparities in subgroups according to wealth. There is a strong relationship\nbetween poverty and chronic undernutrition in both urban and rural areas; this relationship does not change simply by living in an urban environment. However,\nurban and rural living conditions and lifestyles differ, and it is important to consider these differences when designing programmes and policies to address undernutrition.