In wireless sensor networks, the limited battery resources of sensor nodes have a direct impact on network lifetime. To improve energy efficiency, there have been a variety of sensor node scheduling methods to keep only a minimum number of sensor nodes in active mode. However, those existing sensor node scheduling methods mainly focus on the coverage problems, largely ignoring the fault-tolerance issues. In this paper, we proposes a new fault-tolerant sensor node scheduling algorithm, called FANS (Faulttolerant Adaptive Node Scheduling), that does consider not only sensing coverage but also sensing level. The proposed FANS algorithm enables to preserve sensing coverage, maximize network lifetime and achieve energy efficiency. Experimental results show that the FANS algorithm provides enhanced performance in terms of sensing coverage, energy consumption and the number of control messages.