Available results about energy and GreenHouse Gases (GHG) balances of biofuels from Life- Cycle Assessment (LCA) or life-cycle based studies present large discrepancies and thus, may lead to contradictory policy-making measures. This work reviewed seven important European LCA studies in a sensitivity analysis approach in order to get a better understanding of the roots of such a debate for three major biofuels in European production: rape methyl ester and ethanol from wheat and sugar beet. Global trends and variability of energy and GHG balances were depicted and completed with a sensitivity analysis carried out for each methodological and data parameter, which allowed making recommendations on the carrying out of LCA in a policy-making or a biofuels comparison context. Methodological choices, and especially allocation rule, appeared as key elements for results variation with influences on balances up to 149%; system expansion approach was identified as the most relevant rule since it integrates the market potential and the environmental interest of by-products promotion, which was pointed out as a crucial point for biofuels sustainability. The influence of local specificity for cultivation data was evaluated up to 167%, which puts too large geographical coverage in question. Modelling uncertainties due to N2O emissions from soils showed influences from 17 to 46%, which represents a crucial challenge for research and for LCA results accuracy. Approximations evaluation pointed out the need to integrate agricultural machinery into the assessment. Finally, land-use issue revealed its dramatic importance for LCA results and the need to define explicit scenarios for land-use alternatives.