Accurate prediction of wave-induced motions is essential for the design and safe operation of marine and offshore structures. Among these motions, heave and pitch significantly affect structural performance and operational safety. Conventional hydrodynamic methods are often computationally expensive and unsuitable for real-time applications. This study presents a prediction framework based on a Generalized Feedforward Artificial Neural Network (GFANN) to estimate the heave and pitch responses of a semi-submersible platform under various wave conditions. A hydrodynamic dataset generated using strip theory, covering different vessel speeds and wave headings (0°–180°), was used for training and validation. The proposed model accurately predicts vessel motions under unseen conditions while requiring significantly less computational effort than conventional simulations. The results demonstrate the capability of GFANN to capture nonlinear response behavior and support real-time applications such as operational planning, digital twins, and intelligent monitoring in marine engineering.