The GO methodology, which is a success-oriented system analysis technique, is effective for evaluating the reliability of complex systems with multiple states and time-series. It is widely used in the domain of nuclear and ship industry. However, the GO methodology has some restrictions in modeling and analyzing an intricate system that contains dynamic behavior characteristics, such as function dependency, backup dependency, and load sharing. To enhance both the capacity of the modeling and the scope of applications, we proposed an extended GO methodology in this paper to describe the dependencies of the dynamic behaviors. Integrated with the dynamic Bayesian network (DBN), the dynamic behaviors can be presented in a unified way. By using mature software, the extended GO methodology proposed in this paper can be calculated conveniently. Meanwhile, based on unified rules, the multi-operator can be mapped into the DBN, followed by a complete GO model with complex characteristics that can be converted into an isomorphic DBN and analyzed easily by utilizing DBN's powerful inference capabilities. Moreover, the approach makes the extended GO model easy to analyze and intuitive for nonexperts.