Abstract: The reliability of wind turbines is influenced by both natural degradation and external shocks. The Markov chain method is often used to model the natural degradation process of components. The existing related studies ignore the impact of external shocks on degradation, therefore, an improved Markov chain method is used to model the degradation process of components. An improved transition probability matrix is also proposed to describe the influence of external shocks on degradation, and a homogeneous Poisson process is used to describe the random shock process. On this basis, component reliability modeling is conducted, while a revenue model of preventive maintenance strategies is established with the preventive maintenance time being studied. Finally, using a key component of a wind turbine, i.e. the generator, as an example, the effectiveness of the proposed Markov chain modeling method, which considers both natural degradation and external shocks, as well as the economy of the preventive maintenance strategy is verified.