We report global distributions of marine isoprene flux, whose source is estimated by combining an empirical relationship for isoprene production rate with MODIS satellite chlorophyll observations from 2001. We use a steady-state water column model including losses to chemistry, bacteria, and air-sea exchange. Physical mixing is a negligible sink. Flux estimates range from 107--109 molecules cm-2s-1, with considerable spatial and temporal variability, resulting in a global annual total of 0.1 Tg C/yr. Air-sea exchange is the dominant isoprene sink in the surface oceans, with bacteria the second largest, but much less important, sink. The reported fluxes represent a small loss of OH in the remote marine boundary layer (MBL) compared to other oxidants. Application of our approach to other reactive compounds may improve a priori flux estimates for coupled atmosphere-ocean biogeochemistry inverse model studies.