Different ecosystem values of the Amazon rainforest are surveyed in economic terms. Spatial rainforest valuation is crucial for good forest management, such as where to put the most effort to stop illegal logging and forest fires, and which areas to designate as new nationally protected areas. Three classes of economic value are identified, according to who does the valuation: values accruing to the local and regional populations (of South America); carbon values (which are global); and other global (noncarbon) values. Only the first two classes are discussed. Three types of value are separated according to ecosystem service delivered from the rainforest: provisioning services; supporting and regulating services; and cultural and other human services. Net values of provisioning services, including reduced impact logging and various non-timber forest products, are well documented for the entire Brazilian Amazon at a spatially detailed scale and amount to at least $20–50/ha/year. Less-detailed information exists about values of fish, game, and bioprospecting from the Amazon, although their total values can be shown to be sizable. Many supporting and regulating services are harder to value economically, in particular climate regulation and watershed and erosion protection. Impacts of changed rainfall when Amazon rainforest is lost have been valued at detailed scale, but with relative model values of $10–20/ha/year. Carbon values are much larger, at a carbon price of $30/ton CO2, around $14,000/ha as capitalized value. The average per-hectare value of tourism and the health benefits from having the Amazon forest are low, and such values cannot easily be pinned down to individual areas of the Amazon. Finally, the biodiversity values of the Amazon, as accruing to the local and regional population, seem to be small based on recent stated-preference work in Brazil. Most of the values related to biodiversity are likely to be global and may. in principle, be very large, but the global components are not valued here. The concept of value is discussed, and a marginal valuation concept (practically useful for policy) is favored as opposed to an average or total valuation. Marginal value can be below average value (as is likely for biodiversity and tourism), but can also in some contexts be higher. This can occur where losing forest at a local scale increases the prevalence of forest fires and where it increases forest dryness, leading to a multiplier process whereby more forest is lost. While strides have recently been made to improve rainforest valuation at both micro- and macroscales, much work still remains.