Wireless Sensor Networks (WSN) is built with miniature sensor nodes (SN) which are deployed into the geographical location being sensed to monitor environmental condition which transfer the sensed physical information to the base station for further processing. The sensor nodes frequently experience node failure as a result of their hostile deployment and resource limitations. In WSN, node failure can cause a number of issues, namely Wireless Sensor Networks topology changes, broken communications links, disconnected portions of the network, and data transmission errors. An important concern of WSN is the detecting, diagnosing and recovering of sensor node failures. In the course of this effort, an effective strategy for sensor node failure detection algorithm using Poisson Hidden Markov Model (PHMM) and the Fuzzy based Chicken Swarm Optimization (F-CSO) is proposed for efficient detection of sensor nodes fault in the WSN. The proposed work offers optimal false alarm, false positive, energy consumption, detection accuracy, network lifetime, and least delay rates. Moreover, the F-CSO provides improved localization to locate the defective sensor nodes which are present in the WSN. The proposed work is implemented in the NS2 simulator with realistic simulation parameters and the simulation result demonstrate that the proposed work is more effective in terms of false alarm rate, false positive rate, detection accuracy, delay, energy consumption and network lifetime when it is compared with other existing state of art systems.