Technical Abstract: Simulation models for plant species important for biofuel such as switchgrass (Panicum virgatum L.) can be used to make management decisions related to biomass productivity and related to environmental impacts such as soil erosion and changes in surface and groundwater quality. The present study was designed to evaluate the accuracy of simulation of switchgrass biomass production by the ALMANAC (Agricultural Land Management Alternatives with Numerical Assessment Criteria) model at sites in Texas, Arkansas, and Louisiana. We used multi-year field data for Alamo switchgrass yields at each of five sites to evaluate ALMANAC. The model realistically simulated mean switchgrass yields at each of the locations, but did not perform as well in accounting for the year-to-year variability. Sensitivity analysis showed that changes in runoff curve number (CN) and changes in maximum stomatal conductance (GSI) had variable impacts on simulated values among the sites. A 15% change in CN changed mean annual biomass yield from 0 to 16% depending on location. Changing GSI from 0.004 to 0.008 m s**-1 changed mean annual biomass 1 to 31% depending on location. ALMANAC shows promise as a tool to realistically simulate mean biomass yields and variability around the mean for multiyear runs of switchgrass at these diverse sites in Texas, Arkansas, and Louisiana. Further research, with more extensive measurements of soil parameters including soil nutrients is needed to determine why the model reasonably simulated individual years’ yields at Stephenville and Dallas, but had difficulty at other sites.