This study describes a reliability test method for reliability evaluation of MMICs (Monolithic Microwave Integrated Circuits) in product inspection applications. It takes advantages of the potentiality of various reliability test approaches, aiming at meeting the requirement of MMIC development. In this way, quicker realistic reliability assessment can be also realized for new products or those without historic data. Applications of this prediction model to real MMICs are illustrated and a general overview of the corresponding parameters influence is given. The results of this study indicate in order to predict the GaAs MMICs reliability in a fixed shorter time and smaller sample size, one can design the test based on the combination of empirical methods and statistical methods. This study proposed a reliability prediction combing Arrhenius method and Weibull statistical method and we find Weibull slope is important for the MMICs reliability characterization. The analysis predicts excellent reliability for MMICs based on Arrhenius method, Weibulll method and zero fails result.