•A dual-dimensional functional network model based on Bayesian network was built.•A rapid assessment framework for seismic resilience of substations was proposed.•A biological heuristic multi-objective optimization algorithm was presented.•Optimal post-earthquake recover strategies were obtained through multi-level analysis. Electrical substations need a sufficient amount of time to repair damaged equipment and restore power after an earthquake. Yet, the devastation and danger inherent in an earthquake requires fastest return to power possible; thus, finding better methods to improve the efficiency of post-earthquake emergency recover is an urgent issue. This paper presents a rapid seismic resilience assessment framework which combines a network model and functional time-varying feature. The authors developed a dual-dimensional functional network model of a typical 220 kV substation built with an emphasis on its connectivity capabilities and the power transmission capacity of its equipment. The model’s rapid function status was evaluated based on its network dependence on Bayesian network nodes. The authors’ post-earthquake iterative analysis focuses on resource constraint and power user importance. This article shows how the authors obtained the stepped functional time-varying function as a basis for quantification in the post-earthquake recovery process and provides a seismic resilience analysis of the electrical substation. The multi-objective heuristic optimization algorithm was developed to determine an optimal post-earthquake multi-level repair strategy for substation post-earthquake recovery and to determine substation’s seismic resilience levels.