We describe the magnitude of fluctuations in trout populations in several widely separated streams in the intermountain region of the western United States, and consider the potential effect of such fluctuations on land‐management planning. Trout populations included native and exotic species, self‐reproducing and hatchery‐maintained populations, and assemblages that ranged from monospecific to diverse. Annual fluctuations in population statistics were generally large, and some fluctuations were related to geographic setting and trout species. For individual species, fluctuations in all statistics were typically less in the Rocky Mountain study areas than in the Great Basin, but, for the total salmonid community, the situation was reversed. Numerical population fluctuations frequently did not parallel fluctuations in biomass. Except in cases of irregular occurrence, populations of brook trout Salvelinus fontinalis, particularly those in Rocky Mountain study areas, were numerically the most stable; those of allopatric cutthroat trout Salmo clarki in the Great Basin were the least stable numerically. However, biomass of allopatric cutthroat trout was one of the most stable population statistics, and biomass fluctuations were greater for Rocky Mountain brook trout than for most other species. Allopatry and sympatry were not obviously related to species‐specific fluctuations, though there was some tendency for total salmonid fluctuations in number to be lower, and changes in biomass to be higher, in diverse assemblages. In all cases where a species occurred sporadically or regularly but as a minor member of the local assemblage, fluctuations were typically large. The total salmonid community tended to fluctuate less than individual populations, except when fry of anadromous Chinook salmon Oncorhynchus tshawytscha were present. It is apparent that inherent trout population fluctuations must be considered within the framework of land‐use planning if fishery goals are also going to be achieved. Habitat‐based models to evaluate the effects of land uses and habitat enhancement efforts frequently fail to incorporate these fluctuations. For this reason, we concluded that such models often have little utility in predicting sizes or biomass of salmonid populations in the intermountain west.