Minimising the environmental impacts of biofuel production is an urgent global challenge. Over the next decade, increased demand for sugarcane-based ethanol in Brazil could result in over one million hectares of the nation’s native forest and grassland being replaced directly by sugarcane or indirectly by displaced crops and pastureland. Here we integrate future ethanol demand scenarios in Brazil within a spatially-explicit planning framework aimed at minimising impacts of ethanol-driven agricultural expansion on aboveground carbon stocks and 453 species of immediate conservation concern. We show that ethanol-driven agricultural expansion that is blind to carbon and biodiversity values would release 44.9 million tonnes of CO2 equivalent (MtCO2eq), and would impact habitat for at least 273 species. When compared to this conservation-blind scenario, agricultural expansion that avoids carbon and biodiversity values would reduce emissions by 87% (5.8 MtCO2eq) and would avoid impacts on habitat for 113 species. These findings are immediately relevant to policy makers seeking to guide ethanol-driven land-use change away from important environmental areas in Brazil. Our planning methodology can also be extended to other natural areas at risk of bioenergy-driven agricultural expansion.