Published estimates of the potential of bioenergy vary widely, mainly due to the heterogeneity of methodologies, assumptions and datasets employed. These discrepancies are confusing for policy and it is thus important to have scientific clarity on the basis of the assessment outcomes. Such clear insights can enable harmonisation of the different assessments. This review explores current state of the art approaches and methodologies used in bioenergy assessments, and identifies key elements that are critical determinants of bioenergy potentials. We apply the lessons learnt from the review exercise to compare and harmonise a selected set of country based bioenergy potential studies, and provide recommendations for conducting more comprehensive assessments. Depending on scenario assumptions, the harmonised technical biomass potential estimates up to 2030 in the selected countries range from 5,2 to 27.3 EJ in China, 1.1 to 18.8 EJ in India, 2.0 to 10.9 EJ in Indonesia, 1.6 to 7.0 EJ in Mozambique and 9.3 to 23.5 EJ in the US. From the review, we observed that generally, current studies do not cover all the basic (sustainability) elements expected in an ideal bioenergy assessment and there are marked differences in the level of parametric detail and methodological transparency between studies. Land availability and suitability lack spatial detail and especially degraded and marginal lands are poorly evaluated. Competition for water resources is hardly taken into account and biomass yields are based mostly on crude ecological zoning criteria. A few studies take into account improvements in management of agricultural and forestry production systems, but the underlying assumptions are hardly discussed. Competition for biomass resources among the various applications is crudely analysed in most studies and key assumptions such as demographic dynamics, biodiversity protection criteria, etc. are not explicitly discussed. To facilitate more comprehensive bioenergy assessments, we recommend an integrated analytical framework that includes all the key factors, employs high resolution geo-referenced datasets and accounts for potential feedback effects.