For nearly 30 years, technologies for more sustainable land use have been developed and promoted in Rwanda. However, these technologies have not been fully adopted. Keeping in mind that the farming population is not homogeneous with respect to socio-economic variables, this paper typifies farm households in Umutara province based on socio-economic factors influencing the adoption of new technology. A multivariate analysis approach that combines Principal Component Analysis and Cluster Analysis allowed us to identify clearly five types of farm households and their socio-economic characteristics. The main differences between the five farm types relate to gender, age, education, risk perception, risk attitude, labour availability, land tenure and income. The five farm types are characterized by respectively having a female head (26% of the farms), being a tenant (7%), having a male and literate head (32%), having an illiterate head with no off-farm activities (18%), and being a large farm with livestock (17%). The respective farm types appeared to have adopted different types of sustainable technologies to a limited extent. Female-headed households adopted the use of compost and green manure. Young male literate farmers were the only ones using chemical fertilizers. Illiterate and full-time farmers applied fallow, manure and erosion control measures to maintain soil fertility. The use of improved livestock is adopted by large farms.