In this paper, we explored the possibility of using Genetic Algorithm (GA) being used in Wireless Sensor Networks in general with specific emphasize on Fault tolerance. In Wireless sensor networks, usually sensor and sink nodes are separated by long communication distance and hence to optimize the energy, we are using clustering approach. Here we are employing improved K-means clustering algorithm to form the cluster and GA to find optimal use of sensor nodes and recover from fault as quickly as possible so that target detection won’t be disrupted. This technique is simulated using Matlab software to check energy consumption and lifetime of the network. Based on the simulation results, we concluded that this model shows significant improvement in energy consumption rate and network lifetime than other method such as Traditional clustering or Simulated Annealing.