Predicting the C and N mineralization of straw added to soil is important for forecasting subsequent soil N availability during and between crop growth cycles. The decomposition module of the STICS model, parameterized under optimal conditions, was used to predict straw decomposition in sub-optimal conditions, i.e. when contact between soil and residue was poor (due to large size residues or surface placement) or when mineral N availability was restricted. The data used in the simulations were obtained from published studies of effects of residue size, location and N availability on C and N mineralization from straw under controlled laboratory conditions. We selected studies in which the dynamics of C and N mineralization were measured simultaneously. The dynamics of straw mineralization could be well predicted by the model under optimal conditions with standard parameter values as derived from measured C/N ratios of the residues, but not under sub-optimal conditions which required a new parameterization. A good fit could be obtained on these treatments by a marked reduction in the rate constants of residue and microbial biomass decomposition and a marked increase in the microbial biomass C/N ratio. Our results show the need to include in decomposition models routines for simulating effects of spatial heterogeneity of residue distribution, different particle sizes and limiting N availability.