 The solar photovoltaic, PV, energy has become increasingly popular as a source of renewable energy. To better manage electrical systems that use PV arrays, researchers have focused on developing accurate models for predicting solar radiation. Artificial neural networks have proven effective in this area, but existing models do not always meet the needs of certain applications. This paper proposes a new model based on a nonlinear autoregressive exogenous, NRX, neural network that can accurately predict solar radiation levels on a horizontal surface. The model takes into consideration all the specific conditions of a sailboat's operation, including the angle of incidence, the wind speed, and the direction of the sun. The results demonstrate that the NRX neural network provides the most accurate predictions when it is trained periodically. This article was authored by Zina Basada, Octavian Curia, AUKMed RemiSci, and others.