A method of analyzing origin-destination (O-D) connectivity reliability of real-world lifeline networks is proposed in this research. This research is novel because, for real-world lifeline networks with thousands of vertices and links, evaluating O-D connectivity is computationally infeasible with traditional methods, for example, combinatorial methods. It is postulated that the O-Dconnectivity problem for the actual network can be converted into a virtual problem, while the computation for this problem is fast even for large networks. Hence, it is applicable to real-world lifeline networks. When combined with reliability analysis methods, the proposed method can be used to efficiently solve realistic O-D connectivity reliability problems for real-world lifeline networks. Besides, a rating method is also proposed to rank critical links, that is, links that have great influence on the O-D connectivity. The proposed framework should be valuable for management and reliability analysis of real-world lifeline networks.