In this work, a wideband GaN HEMT model with comprehensive hybrid parameter extraction for 5G power amplifier is proposed. Unlike other methods, polynomial fitting is introduced in this paper to achieve more efficient parameter extraction without losing the accuracy of the extraction results. In large signal modeling, artificial neural network is inserted to accurately demonstrate the nonlinear current-voltage relationship in GaN HEMT. The proposed model is built and a 28GHz power amplify is designed based on the model in ADS, the error between the simulation results and the test results is less than 5% by comparison, which completes the model validation.