This paper examined the socio-spatial factors influencing the severity of multi-dimensional poverty at the household level with a view to untangle the variations over space. Studies of multi-dimensional poverty at the household level rarely focus on multiple deprivations, poverty severity and variation in the distribution of multi-dimensional poverty. Stratified sampling technique was adopted while 2521 households were randomly selected for the survey. Geographical information system (GIS), principal component analysis (PCA) and analysis of variance (ANOVA) were used to examine urban poverty variations experienced. Multi-dimensional poverty was measured based on deprivations in more than 50% of multi-dimensional indicators adopted. The most deprived households were found in delta south senatorial district with a severity index of 1729. The severity index of deprived households varies from 0.1729 in Delta South Senatorial District to 0.1463 in Delta North to 0.1418 in Delta Central along education, health and living standard dimensions. The result of the ANOVA based on further analysis of the extracted components in the PCA is statistically significant at p ≤ 0.001 with F value of (F = 527.305). Identification of different poverty categories and inclusion of the multi-dimensionally poor in poverty reduction policies is vital to the eradication of poverty in Delta state.