Flight control system is typical safety critical system, and the reliability of flight control system plays an important role in ensuring the safe operation of aircraft. Traditional reliability analysis methods have a heavy reliance on the experience of analysts, which makes it easy for inconsistencies between reliability models and design models. The fault propagation behavior of the system is thoroughly described by the Architecture Analysis Design Language (AADL) and stochastic Petri nets (SPN), and a method for model-based reliability analysis is proposed. The nominal model and error model of a lateral fly-by-wire flight control system were constructed using AADL. A method for extracting error propagation information from the AADL model was proposed, and the SPN model described the fault propagation behavior of the system was automatically generated by extracting the information of AADL model. Based on the SPN model, Monte Carlo simulation was used to evaluate the reliability of the lateral fly-by-wire flight control system, compared with the fault tree analysis method, the error is less than 0.018%, which can be neglected in practice. Through the method of this study, the reliability model is automatically generated by the AADL model, which ensures the consistency between the reliability model and the design model and avoids reliance on the experience of designers.