A novel neural network (NN) based scheme performs nonlinear Model Output Statistics (MOS) for generating precipitation forecasts from a Numerical Weather Prediction (NWP) model output. The scheme requires minimal past data to establish the MOS connection, and adapts itself to the temporal changes in the NWP model. The method is demonstrated in three numerical experiments using the NCEP reanalysis data in the Alaskan peninsula and the coastal region of British Columbia. Its performance is compared to that of a conventional NN-based non-adaptive scheme. When the new adaptive method is employed, the degradation in the precipitation forecast skills due to changes in the NWP model is small, and is much less than the degradation for the conventional non-adaptive scheme.