We have used a marine food-web model, an atmosphere-ocean general circulation model (GCM), and an empirical dimethylsulfide (DMS) algorithm to predict the DMS seawater concentration and the DMS sea-to-air flux in 10¿ latitude bands from 70¿N to 70¿S under contemporary and enhanced greenhouse conditions. The DMS empirical algorithm utilizes the food-web model predictions of surface chlorophyll and the GCM's simulation of oceanic mixed layer depth. The food-web model was first calibrated to contemporary climate conditions using satellite-derived chlorophyll data and meteorological forcings. For the climate change simulations, the meteorological forcings were derived from a transient simulation of the CSIRO Mark 2 GCM, using the IPCC/IS92a radiative forcing scenario to the period of equivalent CO2 tripling (2080). The globally integrated DMS flux perturbation is predicted to be +14%; however, we found strong latitudinal variation in the perturbation. The greatest perturbation to DMS flux is simulated at high latitudes in both hemispheres, with little change predicted in the tropics and sub-tropics. The largest change in annual integrated flux (+106%) is simulated in the Southern Hemisphere between 50¿S and 60¿S. At this latitude, the DMS flux perturbation is most influenced by the GCM-simulated changes in the mixed layer depth. The results indicate that future increases in stratification in the polar oceans will play a critical role in the DMS cycle and climate change.