We implemented a process-based DMS module into the global carbon cycle ocean model (HAMOCC5) which includes a simple module for plankton dynamics and investigated the regional and seasonal variations of the marine sulfur cycle. The turnover rates within the DMS cycle are only poorly known. Therefore we developed, on the basis of a global DMS data set, an optimization routine for the free parameters controlling DMS production and removal. The resulting seasonal and regional distributions of DMS concentration are fully consistent with the underlying hydrodynamical and biogeochemical processes. We investigated a series of DMS model approaches with various complexities. The distinction between different DMS producing phytoplankton species and the consideration of the regionally and seasonally varying bacterial activity on converting dDMSP to DMS and on DMS consumption appears to have a crucial effect on the quality of the results in the given model conception.