This paper analyzes the energy selection behavior of Nepalese households, specifically the effects of socio-demographic factors on the energy mix. To overcome the limitations of prior research and to accurately model household energy selection and consumption behavior simultaneously, this study uses the multiple discrete continuous extreme value (MDCEV) model. The empirical analysis was performed on data from the Annual Household Survey (AHS) conducted by the Central Bureau of Statistics (CBS) of Nepal. The results show that lower education levels of the household and private house ownership are linked to the consumption of fuels such as firewood and kerosene, whereas ownership of information and communications technology (ICT) devices and access to renewable energy lead to a preferred selection of more modern, cleaner fuels. The results provide implications for policy measures supporting a shift towards the use of cleaner fuels such as LPG and electricity to overcome the negative effects of traditional fuels on personal health and the environment. (C) 2018 Elsevier Ltd. All rights reserved.