An econometric model linking user household’s socio-economic characteristics and their fuel-wood collection patterns from local common property forests was examined in a highly heterogeneous (>60% by caste and wealth status) Terai community of Nepal. The cross-sectional analysis provides evidence in favor of the hypothesis that socio-economic inequalities within the group are inexorably associated with the ability of the households in resource use. In general, it is evident that household’s wealth status (coefficient for rich = -0.420; p = 0.019), proximity to the forest (coefficient for distance = -0.280; p = 0.083), forest visit (coefficient for frequent forest visit = +0.257; p = 0.066) exert a strong influence on appropriating fuel-wood from the forest. Above all, income status of households was found to be key determinant of household’s fuel-wood collection from the forest. Poor households were highly dependent on the forests for fuel-wood (average annual extraction = 4561.3 kg/household) in order to sustain their day-to-day livelihood. The high dependence of poor coupled with their large population size in the region (>27%) will possibly cause forest degradation in future. Therefore, policy must be directed towards reducing fuel-wood dependency of poor households by lifting their economic status so that they can substitute fuel-wood in day-to-day use. A broader policy implication of our analysis is to consider socio-economic heterogeneity of society and household’s dependency on natural resources prior to devolution of management rights to local level.