Emissions of trace gases and particulate matter from burning of biomass are generally factored into global climate models. Models for improving the estimates of the global annual release of emissions from biomass fires are presented. Estimates of total biomass consumed on a global basis range from 2 to 10 Pg (1 petagram = 10^6 g) per year. New data are now available for emission factors for a range of vegetation types and burning conditions for the tropical, mid-latitude, and boreal forest regions. We estimate the total emissions of CO, CO2, CH4, nonmethane hydrocarbons (NMHC), and particulate matter with a mean mass cut-point diameter of 2.5 um (PM2.5) based on our algorithms for emission factors for different fuel type and burning condition variables. These emission factors are dependent on the efficiency of the fire in oxidizing fuel carbon to CO2--combustion efficiency (CE). For flaming combustion, the CE is high. On the contrary, for smoldering combustion, the CE is low and emission factors for PM2.5, NMHC, CH4, and CO are high. With the exception of emission factors for CO2, the emission factors for tropical savannah burning are lower than those used for many global estimates of emissions from biomass burning. Our total global estimates of emissions for PM2.5, NMHC, CH4, and CO are 48.7, 21.1, 28.4, and 362 Tg per year (1 Teragram = 10^2 g). One of the advantages of using algorithms based on a CE argument is that the algorithms are dynamic in that they can account for the changing effect of climate and cultural practices on fire behavior and emissions production.