This study investigates seven dimensions of poverty in Vietnam (income, health, education, housing, assets, basic services and economic status) using the Household Living Standard Survey data of 2014. The Government of Vietnam disburses funds for poverty alleviation to regions on the basis of incidence of household income poverty. Our study shows that this method neither fully captures the complex regional diversity of poverty nor does it accurately identify regions with a higher severity of poverty. For the first time in poverty studies of Vietnam, we explore the role of multiple spatial levels on poverty in multiple dimensions. Unlike the practice in the existing literature which classifies the poor with an arbitrary poverty cut-off, we use a fuzzy method that allows the inclusion of people who are in partial poverty. Furthermore, by utilizing random intercept multilevel models to decompose the variation of poverty at the household, commune, district and province levels, poverty maps for Vietnam are developed to visualize the spatial evidence of the severity and incidence of poverty. We identify that the provinces that are relatively less (more) poor in the income dimension are more (less) destitute in several other dimensions, which clearly shows a need for special policy attention. Our method reveals that the poverty ranking of provinces in regional Vietnam departs widely from those obtained through traditional single-level analysis. This suggests that poverty in Vietnam can be explained not only by characteristics at the household level, but also by contextual factors at higher levels (commune/village, district, province). These empirical findings can help Vietnamese policy makers determine suitable strategies to effectively target the most deprived regions and to develop more appropriate poverty-alleviation programs.