Despite the intrinsic relationships between the natural environment, poverty and human wellbeing, measures of poverty do not adequately account for environmental dimensions of wellbeing. This paper furthers theoretical debates about environment-poverty relationships by developing environmental dimensions and indicators for integration into multidimensional poverty indices. We demonstrate that this integration is practically possible using public datasets that are part of regular data collection efforts. Using Brazil as a case study, we develop an environmentally-adjusted multidimensional poverty measure that combines data on proximity to natural land, floods, droughts and landslide risks with information on health, education and living standards. Our results show that the integration of environmental dimensions into a poverty measure can provide different estimates of the incidence and distribution of poverty in a country, shifting attention to new target areas for poverty reduction strategies. Our measure provides a complementary and broader understanding of poverty than traditional poverty indices and could be used to track changes over time. Our measure therefore provides a potentially useful avenue to assess progress towards achieving multiple Sustainable Development Goals, while also accounting for environmental change.