This study aims to see the impact of variables on poverty from the income measure. This will employ a probabilistic model (probit). The variables used represent characteristics of persons, and home environment in which they live. This will be done for the Colombian from the National Household Survey of the year 2000 by the DANE. The results show that environmental variables (such as city, independent work and temporary work) are the most bearing for a person to be poor. In addition to these variables such as level of education of persons, the number of children and if you work for the government indicate the incidence of people to be poor.