Agricultural practices and land use significantly influence soil carbon storage. The processes that are affected by land use and management are generally understood, but uncertainties in projections are high. In this paper, we investigate the long-term effects of chronosequential land use change from grassland to cropland and vice versa on soil carbon stock dynamics in four fields on a Swedish farm. Between 1850 and 1920, three of the fields were converted from grassland into cropland, and one was converted back to grassland in 1971. The fourth (control) field is a grassland that has never been ploughed. In 1937, the four fields were sampled at 111 points in a regular grid (25 or 50 m) and the dried soil samples were stored at our Department. In 1971 and 2002, the original grid points were revisited and re-sampled. Land use changes affected the soil C stock significantly. In 1937, carbon stocks were significantly smaller in the arable fields than in the grassland soil. In the field that was converted from arable back to grassland, soil C increased significantly at an average rate of about 0.4 Mg ha−1 year−1. A soil C balance model (ICBM) driven by standard meteorological data and soil carbon input estimated from yield records described soil carbon dynamics reasonably well, although the range of simulated relative changes in C stocks between 1937 and 2002 in the four fields (from −7.4 to +8.8%) was narrower than those measured (from −19.5 to +16.5%). There are only few long-term studies in Northern Europe available for quantifying the effect of land use change on soil carbon stocks and the results presented here are therefore useful for improving predictions of changes in soil carbon driven by land use change.