The flux of isoprene, one of the more reactive biogenic volatile organic compounds, was measured using eddy covariance techniques on a continuous basis during the 2000–2002 growing seasons at a mixed hardwood forest in northern lower MI. Daytime fluxes of isoprene and both sensible (H) and latent heat flux (LE) were linearly correlated with a positive slope on a daily basis, but the slopes of these relationships varied from one day to the next. Under drought conditions and longer time periods, LE fluxes were suppressed yet isoprene emissions were enhanced; thus, the slope of the isoprene to LE relationship was increased. For example, seasonally averaged LE fluxes were reduced in 2000 as a result of reduced rainfall, and average isoprene fluxes were 1.5 times greater in 2000 compared to years 2001 and 2002. The strong daily correlation between isoprene fluxes and associated energy fluxes is an important relationship that should be accurately reflected in canopy models used for estimating biogenic emissions. In addition, in cases where land surface model output includes latent heat fluxes, the latent heat fluxes may be used as a surrogate for above canopy light and temperature to provide a simple estimate of isoprene emissions. Observed isoprene fluxes are compared to the Biogenic Emission Inventory System (BEIS3) biogenic emission model and to a canopy scale biogenic emission model (WSU-BEIS) that includes a leaf energy budget and accounts for vertical changes in light and temperature through the canopy. Comparison of the emission models with observations showed that the fractional gross errors in isoprene flux estimates were approximately 37% for both models in 2001 and 2002, but increased to approximately 62% in 2000 when less rain occurred and LE fluxes were reduced. The BEIS3 model does not treat sensible or latent heat flux, but the simple scaling and leaf energy budget used in WSU-BEIS yields estimates of LE within approximately 50% of observations. Estimates of sensible heat flux with WSU-BEIS were larger than observed and indicate that the simple scaling approach may not be capable of treating all of the dynamics of canopy energy transfer accurately. Nonetheless, the strong relationships observed between isoprene flux and the energy flux terms provide a basis for testing these and other canopy models used for estimates of biogenic emissions.