Temperature effects are considered in compact modeling gate-all-around (GAA) nanosheet (NS) MOSFETs based on a fundamental device current equation and two different artificial neural networks (ANNs) from -75°C to 125°C. The first ANN, constructed with the drain current model, so-called the Grove-Frohman (GF) model, captures the main trend of device I-V characteristics and predicts variations in device parameters. The second ANN generates correction terms for improving the accuracy of drain current. Applying this deep-learning model to various circuit simulations and comparing it with raw data cross different temperatures consistently yield excellent results.