Preparing regional development strategies for the Amazon Basin is a vexing task for policymakers. Forests continue to fall and agriculture to move in to a region with patchy (in terms of agronomic potential) yet broadly nutrient-poor soils. The spatial distribution of soil types is not well mapped at finer scales relevant for agriculture. There is, moreover, little evidence about how farm land use or farm household welfare varies by soil quality in this frontier setting. Despite these information gaps, regional planners continue to use soils as a basis for policy action, some of which may influence future options for the Amazon. This paper uses a farm-level bioeconomic model that captures soil-quality-specific degrading effects of agricultural activities to assess the impacts of soil quality differences on deforestation, use of cleared land, and smallholder income in the western Brazilian Amazon. Focusing on an archetypical area farm with reasonable market access but limited access to labor and credit, simulations show soil quality mattered more for income than for deforestation or land use, although extremely cash-strapped farmers on poor-quality soils could face displacement. Pasture dominated farm land use across all soil types as farm forest disappeared within a generation on successfully established farms. Good- and (viable) poor-soil farms had slightly slower deforestation rates than their medium-soil counterparts – rich-soil farms shifting small amounts of area (and labor) to the more nutrient- and labor-intensive annuals, and (even viable) poor-soil farms lacking sufficient resources to clear and farm additional land. Farms with good soils could generate about 44 per cent more income than their viable poor-soil counterparts, but the lower-income level still surpassed thresholds for meeting food security and other needs. At no combination of income level and soil quality explored did the (simulated) farmer find it worthwhile to purchase and apply chemical fertilizer; nutrients came instead from secondary forest fallow, whose area rose or fell in step with annual cropping area. The implications of these results for land-use zoning, forest conservation, poverty alleviation, and other policies are discussed.