Software fault prediction is the most efficient methodology to improve the quality of the products. To improve the quality it is essential to find the error or fault as quick as possible. To determine and advance the development of product there are different prediction approaches are available like correction cost prediction, test effort prediction, software fault prediction (SFP), security prediction and so on. This paper survey the variety of methods and metrics used to improve the quality of the object oriented software. We compared different prediction models and proposed the methodology with the datasets they used for easy understanding. Software fault prediction using different techniques to improve the quality and the error free software delivery. This paper gives overview about the prediction models for software in design and implementation phase and also it discuss trends which are currently used software fault prediction.