A large sensor network needs multiple base stations(BSs) to collect data efficiently.The design effort often aims at maximizing the network lifetime.The lifetime of a network depends on,besides battery capacity,factors such as node density,sampling period and the partitioning of the network for the multiple base stations.In this paper,we propose a stochastic approach to finding the lifetime of large sensor networks with multiple BSs and randomly scattered nodes.We begin by partitioning the network into regions by the Voronoi diagram method,one for each base station.We then propose a method to find the minimum transmission range needed for guaranteeing network connectivity.With nodes connected,we derive the traffic rate of nodes at specific distance from the base station.Based on the traffic rate and energy consumption model,the lifetime distribution of a region is calculated.Combining the results of individual regions,the network lifetime distribution is obtained.Case studies are presented and verified by computer simulation.