In heterogeneous wireless sensor networks, a strong relation exists among different types of data being composed into an event. Energy consumption can be reduced by collecting partial data to approximately detect the event. A node scheduling problem for approximate event detection is proposed. The nodes in the network are organized into subsets that are activated successively. The lifetime of the network is maximized on the condition that the partial data collected can be formed as an approximation of the event. The node scheduling problem for approximate event detection is proved to be NP-complete. A greedy-based approximate algorithm is devised with 1/2 as the approximate ratio. The experimental results illustrate that the lifetime of the network can be prolonged effectively.