This study investigates different pipe recovery strategies of water distribution networks (WDNs) from the seismic resilience perspective. A demand-based seismic resilience index (SRI) is defined to reflect the system performance level during the post-earthquake recovery process, which is characterized by the satisfaction degree of consumers. Then, a pipe fragility model based on the Chinese code is employed to determine the pipe damage conditions subjected to earthquakes. Taking the WDN in Mianzhu city as an example, three pipe recovery strategies, i.e., static importance-based strategy, damage-based strategy, and distance-based strategy are compared using the Monte Carlo simulation. After comparing the recovery trajectory of system satisfaction degree and the final SRI, the static importance-based strategy is proved to be the best among the three, especially in the early recovery stage. In addition, the dynamic importance-based strategy and the optimal strategy obtained by the genetic algorithm are discussed in a deterministic way. Results show that although the two strategies improve the seismic resilience of WDNs, the computation cost is high due to the time-consuming nonlinear hydraulic analysis. Compared with them, the static importance-based strategy is still a good one in terms of computational efficiency and high SRI value.