Land use and land cover changes in the Brazilian Amazon region have major implications for regional and even global carbon cycling. We analyzed the effects of the predominant land use change, conversion of tropical forest to pasture, on total soil C and N, using the Century ecosystem model and data collected from the Nova Vida ranch, Western Brazilian Amazon. We estimated equilibrium organic matter levels, plant productivity and residue carbon inputs under native forest conditions, then simulated deforestation following the slash and burn procedure. Soil organic matter dynamics were simulated for pastures established in 1989, 1987, 1983, 1979, 1972, 1951, and 1911. Using input data from the Nova Vida ranch, the Century model predicted that forest clearance and conversion to pasture would cause an initial decline in soil C and N stocks, followed by a slow rise to levels exceeding those under native forest. Simulated soil total C and N levels (2500 g C m−2 and 245 g N m−2 in the 0–20 cm layer) prior to conversion to pasture were close to those measured in the native forest. Simulated above- and below-ground biomass for the forest and pasture were comparable with literature values from this region. The model predicted the long-term changes in soil C and N under pasture inferred from the pasture chronosequence, but there was considerable variation in soil C stocks for pastures <20 years in age. Differences in soil texture between pastures were relatively small and could not account for much of the variability between different pastures of similar ages, in either the measured or simulated data. It is likely that much of the variability in C stocks between pastures of similar ages is related to initial C stocks immediately following deforestation and that this was the largest source of variability in the chronosequence. Internal C cycling processes in Century were evaluated using measurements of microbial biomass and soil δ13C. The relative magnitude and long-term trend in microbial biomass simulated by the model were consistent with measurements. The close fit of simulated to measured values of δ13C over time suggests that the relative loss of forest-derived C and its replacement by pasture-derived C was accurately predicted by the model. After 80 years, almost 90% of the organic matter in the top 20 cm was pasture derived. While our analysis represents a single ‘case study’ of pasture conversion, our results suggest that modeling studies in these pasture systems can help to evaluate the magnitude of impacts on C and N cycling, and determine the effect of management strategies on pasture sustainability.