Molecular property prediction is a research field and computational methodology focused on estimating physicochemical, biological, or other relevant characteristics of molecules directly from their structural representation or computed descriptors. This approach leverages computational models, frequently employing machine learning techniques, to infer properties without requiring experimental synthesis or measurement, thereby accelerating the discovery, design, and optimization processes in chemistry, materials science, and pharmacology.