The development of accurate forecasting methods for renewable energy sources can act as an important tool to integrate renewable power systems in the electricity grid. This paper proposes a technique that can forecast the power production of a photovoltaic plant one day in advance. The procedure is based on a regression model that considers the weather forecasts of the US Global Forecasting Service (GFS) as inputs, and it is trained and tested on a year of power production data of a 1.3 MW plant located in Borkum, Germany. The Multilinear Adaptive Regression Splines method was used to automatically define a reasonably simple model for the system with regression coefficients that could be easily interpreted. The results indicated that the forecasted power obtained by the model exhibited a high correlation with the measured data and relatively low errors despite the limited number of features that were included in the model and a low number of training samples. (C) 2017 Elsevier Ltd. All rights reserved.