Solar yield prediction plays a pivotal role in enhancing electricity generation from photovoltaic (PV) systems and supporting the efficient sizing and management of PV installations. As utility-scale PV deployment continues to grow, accurate forecasting of solar yield is essential for maintaining grid stability and balancing energy production and consumption, especially given the inherent variability of solar power. This study presents an approach for predicting solar yield using Artificial Neural Networks (ANNs), more specifically Feedforward Neural Networks (FNN). We examine multiple FNN architectures to determine the most suitable structure for accurate solar yield forecasting. Our findings show that the more complex FNN architecture delivers superior performance compared to simpler models and significantly outperforms the persistence model in key metrics such as Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and R-squared. These results highlight the effectiveness of advanced ANN configurations in improving prediction accuracy, thus contributing to more reliable and efficient solar energy management.