A model was developed to help resolve the decision of how many fire fighters a large forest fire management agency should hire for a fire season to minimize expected cost plus fire loss. It addresses the use of fire fighters for both initial and extended attack, the temporary hiring of extra fire fighters to satisfy peak demands, and the movement of fire fighters between regions within the agency's jurisdiction. It also recognizes the increase in subsequent fire fighter demand caused by fire fighter shortages. The fire management system is modeled as a Markov chain where the system state is related to the fuel moisture conditions throughout the protected area, and the total number of extended attack fires burning. Tests using representative hypothetical data demonstrated that it is a tractable approach for roughly evaluating fire fighter requirements. The model was used to evaluate the benefits of the centralized control of fire fighters, and indicated that it could reduce expected cost plus loss as much as 30% in the hypothetical example. For. Sci. 34(3):647-661.