The precise modeling of unmanned aerial vehicle (UAV)-integrated photovoltaic (PV) systems is central for their implementation on UAVs, especially due to their continuous movement. Accurate modeling is essential for developing control methods, conducting performance studies, and designing the whole system considering PV system parameters. This original study presents a novel modeling procedure for comprehensive modeling of UAV-integrated PV modules using multi-input multi-output (MIMO) and multi-input single-output (MISO) machine learning models. These models are developed based on a historical environmental dataset and a randomly generated dataset representing flight conditions. Both datasets get processed during the preparation phase. In this phase, two sets of empirically derived group of modified equations (GME) of the PV module are utilized. In addition, the Whale Optimization Algorithm (WOA) is employed to optimize one set of GME containing PV system five unknown parameters. Moreover, the Dogleg Trust Region Algorithm (DTRA) is used as an unconstrained problem solver in each loop of WOA to solve a system of equations consisting of the unknown parameters. Finally, the MIMO and MISO models will be trained with the processed dataset. The proposed approach predicts UAV-integrated PV module behavior under diverse flight conditions, interprets panel movements, and is validated through experiments.