Evidence Based Decision Making using Neural Networks for Software Applications The Bayesian Networks use Conditional Probability Tables for predicting the Cause and Effect Relationship. Bayesian Networks can be replaced by Neural networks which improves the effectiveness of expert knowledge by analysing the data sets. Neural Networks are created by connecting processing elements. Neural Networks replaces the Bayesian Networks because of its ability to solve a problem in software without defining and describing the method of solving such problems and even without any case of personal knowledge about the solved problems. This study focuses on predicting the reliability of any software project depending upon their project characteristics using Neural Networks.