Socio-economic diversity can help to bring about innovative development in agroforestry practices. The diversity of households in the mid-Nepal hills was analysed using survey data from 521 randomly selected households in six villages. A cluster analysis derived the following household typology based on socio-economic variables—Type 1: resource-poor Brahmin/Chhetri; Type 2: resource-poor Janajati; Type 3: resource-rich mixed-caste households; Type 4: resource-rich Brahmin/Chhetri; Type 5: resource-rich Janajati; Type 6: resource-poor Dalit households. The analysis revealed that social status (caste/ethnicity), household status on foreign employment and landholding are strong predictors of household segmentation in rural Nepal. This paper suggests revision of existing wellbeing ranking approaches using these socio-economic variables for more inclusive and equitable agroforestry and community forestry outcomes.