 This paper proposes a new artificial neural network, ANN, model called ANNSFP which uses statistical feature parameters, SFPs, as inputs instead of raw data to improve the accuracy of short-term solar irradiance forecasting, STSIF. The SFPs are derived from the historical data series and include the mean, standard deviation, maximum and minimum values of the irradiance and ambient temperature. The model structure is determined by cross-validation, CV, and the Levenburg-Marquardt algorithm, LMA, is used for the network training. Simulation results show that the forecast accuracy of the proposed ANNSFP model is significantly better than the conventional ANN model using historical data series, ANN-HDS. This article was authored by Hongshan Zhao, Shursu, Xinjiang Mi, and others.