As one of the most critical blocks of RF front-ends for the wireless/mobile communication system, the reliability of power amplifier is crucial for the transceiver. In order to analyze and predict its reliability, accurate modeling and testing are the effective methods to do this. Fast reliability modeling based on artificial neural network is available and high-efficient for PA reliability analysis. Meanwhile, in order to study the index failure of the PA, the accelerated degradation test is carried out to investigate its performance degradation. This paper reports a novel reliability test method and a fast and high-efficient modeling technique to investigate and analyze the PA reliability comprehensively.