In this paper,the max-flow /min-cut algorithm is applied to the event detection in wireless sensor networks,and an event detection algorithm ( G-Cut) is proposed for boundary-abrupt events.In the algorithm,firstly,the sensing data of adjacent nodes are transformed into weight values,and thus a flow network is formed.Then,the event boundary is obtained by using the max-flow /min-cut algorithm to cut the flow network.Finally,the event area is determined according to the direction implied in upload information.Simulation results on wildfire events show that the proposed algorithm achieves a high accuracy of event detection with a small computation amount of nodes,and for multi-event regions,it can guarantee the detection accuracy without increasing the computation amount and traffic of nodes.