This chapter introduces a comprehensive framework for estimating and forecasting harmonics in large, uncertain transmission networks. It leverages advanced machine learning techniques, specifically artificial neural networks, to forecast harmonic distortions at unmonitored buses based on limited PQ monitor installations. The chapter details the development and training of these models, as well as their validation through extensive simulations. Case studies illustrate the model’s ability to accurately estimate harmonics under various conditions, highlighting the benefits of this approach over traditional methods. The potential for applying these models to inform network planning, operation, and compliance with harmonic standards is discussed, emphasizing their value in managing the complexities of modern power systems.