In this paper, a multivariate polynomial (MP) model combined with wavelet analysis is proposed to improve the accuracy and parsimony of 1-month ahead forecasting of monthly anchovy catches in northern Chile. The proposed forecasting model is based on the decomposition the raw data set into low frequency (LF) and high frequency (HF) components by using stationary wavelet transform. In wavelet domain, the LF component and HF component are predicted with a linear autoregressive model and multiscale polynomial autoregressive model; respectively. We find that the proposed forecasting method achieves 99% of the explained variance with reduced parsimony and high accuracy. Besides, the proposed forecaster proves to be more accurate and performs better than the multilayer perceptron neural network model.