Poverty is a multidimensional phenomenon that causes difficulty for people to meet their needs. The research aims to scrutinize physical and social infrastructures concerning multidimensional poverty levels using the spatial approach. Jabung District, Malang Regency, Indonesia has 35% poor households in this case study. The objectives are to measure multidimensional poverty levels, social capital indices of the rate of participation (RoP) and density, and scrutinize neighborhood relationships among 15 villages using spatial regression analysis. Data collection is through a questionnaire survey of 274 heads of households. The research identified four poverty levels (very low to high), where five villages with high poverty levels (Jabung, Taji, Kemiri, Gunungjati, Slamparejo) became the targeted areas. The majority of the villages had a medium level of both the RoP and density, and the community had moderate social relations among community members. The spatial regression analysis indicates that the attribute of the RoP and weight matrix have a significant impact on the poverty level. It is recommended that poverty alleviation programs should focus upon the cluster of poor villages through social infrastructure development as the action to end poverty.JEL Classification A13; I32; R58