A soil community food web model was used to improve the understanding of what factors govern the mineralisation of nutrients and carbon and the decay of dead organic matter. The model derives the rates of C and N mineralisation by organisms by splitting their uptake rate of food resources into a rate at which faeces or prey remains are added to detritus, a rate at which elements are incorporated into biomass, and a rate at which elements are released by organisms as inorganic compounds. The functioning of soil organisms in the mineralisation of C and N was modelled in the soil horizon of a Scots pine forest. The organic horizon was divided into three distinct layers, representing successive stages of decay, i.e. litter, fragmented litter, and humus. Each of the layers had a different, quantitative, biota composition. For each layer the annual C and N mineralisation rates were simulated and compared to observed C and N mineralisation rates from organic matter in stratified litterbags. Simulated C and N mineralisation was relatively close to measured losses of C and N, but the fit was not perfect. Discrepancies between the observed and predicted mineralisation rates are discussed in terms of variation in model parameter values of those organisms that showed the highest contribution to mineralisation rates. The measured, and by the model predicted, significant decrease in mineralisation rates down the profile was not explained by the biomass of the primary decomposers and only partly by the total food web biomass. Modelling results indicated that indirect effects of soil fauna, due to trophic interactions with their resources, are an important explanatory factor. In addition, the analyses suggest that community food web structure is an important factor in the regulation of nutrient mineralisation. The model provided the means to evaluate the contribution of functionally defined groups of organisms, structured in a detrital food web, to losses of C and N from successive decay stages.