Urbanization has threatened rural communities’ livelihoods worldwide, changing their agro-food systems from locally produced traditional items to industrialized foodstuffs. The main objective was to investigate the relationship between livelihood conditions and the agro-food transition process in rural communities of the Center-West, Northeast, and Amazon regions of Brazil. We hypothesized that traditional agroecosystems and local food habits changed with greater access to market economies. The study was conducted with semi-structured questionnaire interviews to verify agro-food patterns, subsistence farming, natural resource use, and socioeconomic conditions. Moreover, we used stable isotope ratios from the inhabitants’ fingernails to determine the food source and trophic chain diversity. Data from questionnaires were analyzed using a Bayesian clustering model to characterize the socioeconomic conditions and agro-food patterns among rural and urban communities. The isotopic data were appraised through a nonparametric model to assess food differences among Brazilian regions and different community types. The Bayesian model allowed us to determine the optimal number of groups according to descriptive socioeconomic and agro-food variables sorted by each specific location. We also verified a food change from C
3
(more natural) to C
4
(more processed) with an increase in
δ
13
C and a decrease in
δ
15
N in the city and town localities. This indicates a livelihood shift from locally produced foods to processed items toward urban areas. Although remote villages showed more maintenance of their agro-food systems, increased access to market economies and the supermarket diet is changing the livelihood conditions of rural communities, which can compromise their traditional farming and food sovereignty.