The relative contributions of double counting of carbon emissions between forest-to-nonforest cover change (FNCC) and forest wildfires are an unknown in estimating net forest carbon exchanges at large scales. This study employed land-cover change maps and forest fire data in the four representative states (Arkansas, California, Minnesota, and Washington) of the US for the period from 1992 to 2006 to evaluate forest carbon double counting effects based on land-cover change map, forest fire data, and USDA Forest Service Forest Inventory and Analysis data. The analyses were conducted at the county level and tallied to state level. Although the effects were small in the two eastern states because of small burned areas and low burn severity, substantial effects were found in the two western states. Carbon double counting was about 10 TgC (teragram 10¹²) in California and 6 TgC in Washington for the period 1992–2006 (at rates of 0.7 and 0.4 TgC per year), or 21.9 and 7.6% relative to total forest carbon emissions through FNCC in the two states, respectively. The effects were 0.2 and 0.1% in Arkansas and Minnesota, respectively. Variation in double counting effects within the states was also much higher in the western states compared with the eastern states. Our results suggested a general pattern that rates and amounts of double counting in forest carbon emissions between FNCC and fires were more evident and substantially different on a west–east dimension than that on a north–south dimension across the conterminous US during the study period.