A study was conducted to assess biomass energy supply and use in two villages in different agro-ecological regions of Bangladesh using three different techniques: broad questionnaire survey, sample questionnaire survey, and physical monitoring in order to see which methodology was more appropriate for the accurate assessment of biomass cooking energy supply and demand. The study included four broad socio-economic groups, i.e. large, medium, small and landless households. II was found that, per capita cooking energy demand varied significantly between the villages, socio-economic groups, and the first two methods of survey. Energy use was found to be strongly correlated with family size and energy use decreased according to landholding size. The broad questionnaire survey showed an average cooking energy use of 11.8 and 10.1 GJ per capita/year, for Kazirshimla and Noagaon village, respectively, while the sample questionnaire survey estimated energy use at 9.2 and 8.5 GJ per capita/year. With physical monitoring, however, the respective figures were found as 9.0 and 8.2 GJ per capita/year. It was also found that energy use was at the subsistence level in both villages. In Kazirshimla village biomass supply was found to be marginally surplus to use, while Noagaon had a supply deficit. The study showed that a broad questionnaire survey gives a fairly general picture of both biomass supply and use, but to assess cooking energy use more accurately successive sample questionnaire surveys followed by physical monitoring are more effective.