Millions of households experience fuel poverty around the world, commonly defined in broad terms in the early 1990s to cover households whose total energy bills exceeded 10% of their net income. First, this paper outlines the measurement of fuel poverty using the Low Income High Costs indicator (LIHC) and examines the main characteristics of fuel poor households in France based on multidimensional statistical analysis using a new micro-level survey data. It then explores the salient determinants of fuel poverty risk using logistic regression. About 3.18 million households are estimated to be in fuel poverty, representing 12% of all French households. The hierarchical classification suggested four distinct profiles, which shed light on the main features of fuel poor households. Additionally, results show that the risk of being fuel poor and the depth of fuel poverty increase significantly with lower EPC rating house and confirm the existence of a social gradient in fuel poverty. Recognition of this aspect can be helpful for developing economically efficient policies to address fuel poverty. This study does not aim to be exhaustive in policy implications terms, but rather to present a new way of thinking about fuel poverty solutions by targeting different household groups.