Based on 1,400 pieces of questionnaires collected in Beijing, this study applies LISA to identify the spatial clustering households with high and low carbon emissions, and explores the relationship between household carbon emissions and urban forms via intensive case studies on five neighborhoods. Results indicate that all of building density, land use pattern, employment accessibility, and public transportation accessibility influence remarkably on household carbon emissions, which suggests that the rational planning of urban form is necessary for the construction of low-carbon cities.