Control of desynchronization in discrete neural networks by nonlinear delayed feedback was studied. The network consisting of Mapbased neurons could get chaotic bursting synchronization through adjusting the coupling strength between the cells. Once nonlinear delayed feedback signals were applied into the fast variables of neurons, a complete desynchronization was achieved and neuron’s inherent bursting characteristic was restored. Compared with linear delayed feedback control, the method was effective even in the case of strong coupling and robust against parameter variations.