The problem of getting an homo-geneous group of elements linked to various kinds of fuzzy relations is presented often in several decision areas. It is possible to solve it by applying the affinity concept that allows grouping items that meet certain characteristics in a degree.
Poverty is a key issue in the analysis of the countries' social situation. This has encouraged several researches in order to quantify this phenomenon, to characterize it and to compare it intertemporally into the country or with other nations.
To target households considered poor due to unsatisfied basic needs it is useful to find an optimal segment, resulting from the crossing of several criteria that identify better the shortcomings of these households to group them by affinity.
In this paper we introduce a linguistic model that uses affinity theory that permits grouping poor households depending on the degree of dissatisfaction of certain basic needs and the importance of the application of public policies to diminish them.