The effect of boundary layer reactions on the conversion of char-N to NO, N2O and HCN was studied by a coupled experimental/theoretical approach. In the experimental part, the evolution of HCN at conditions of a batch fluidized bed reactor was studied. The results show that the char-N conversion to HCN is very low under normal oxidizing conditions. This explains why it was not detected by early studies of this system. The results also suggest that the release of HCN is related to gases released when the char is oxidized rather than to late devolatilization reactions. The product distribution is modeled using heterogeneous reactions on the char surface coupled to homogeneous reactions in the gas phase. Three cases were analyzed. In Case A, only the heterogeneous production of N2O and NO was considered. Case B considered HCN and NO heterogeneous production. Finally, Case C combined cases A and B by allowing the heterogeneous production of NO, N2O and HCN. Cases A and B predict most of the qualitative trends for the char-N conversion to N2O reported in the literature. However, Case A fails to predict the experimentally observed release of HCN when oxidation around the particle is suppressed. Among the three cases, Case C is the one that closely predicts experimental observables. Finally, the predictions obtained by the model are used to define further experiments, particularly on the evolution of gas concentration in the boundary layer with char burnout, that can be useful in order to obtain more insight in the mechanism of nitrogen oxide formation during char oxidation.