Detecting faults in the initial stage of the development process has become an important prospect for the codes cost estimation; so a fault predictor model is very much necessary in order to bring down the cost of development and maintenance. Due to these reasons, developing models for fault prediction has become a crucial part of research, and various techniques have been adapted in order to predict faults in a software. Few of them include Artificial Neural Network, Decision Tree, Genetic Algorithm, etc. Among these techniques, Neural Networks and Genetic Algorithms have become a growing concern over the years and are being applied in various fields such as optimization, prediction or classification. These techniques make use of various software metrics to assess the characteristic of any software system such as number of faults, maintenance of class, etc. Most commonly used are Chidamber and Kemerer (CK) metrics which are found to be efficient from many researches.