A novel algorithm for the digital discrimination of neutrons and gamma-rays in a mixed radiation field is presented. Most of the pulse shape discrimination methods in scintillation detection systems use time-domain features of the signal (e.g. charge comparison method or constant-time discrimination). However, there are no frequency-domain discrimination methods up to date in the literature. Our method employs the wavelet transform to extract frequency-domain features for discrimination. Compared to the pulse gradient analysis (PGA) algorithm, it provides an improvement in reducing the overlap area between neutron and gamma events and also in increasing the Figure of Merit (FoM). Another advantage of this method consists in the removal of the dependency of the discrimination method on timing features. Since the light output in the scintillation process is very noisy, this kind of dependency may degrade the performance of the algorithm.