Data cleansing is a critical step for data preparation. The values lost in the database are a common problem faced by data analysts. Missing values in data mining is continual troubles that can grounds errors in data analysis. Randomly missing elements in the attribute/dataset make data analysis complicated and also confused to consolidated result. It affects the accuracy of the result and intermediate queries. By using statistical / numerical methods, one can recover the missing data and decrease the suspiciousness in the database. The present research gives an applied approach of Newton Forward Interpolation (NFI) method to recover the missing values and other different methods also.Data in the dataset is always remaining as the basic building blocks for any query and further task and decisions. If basis data is incomplete or dataset have missing values the none cannot assume about well up to date final reports. In data mining missing values recognition and recovery is still major issue with irregular data. To overcome from such situation there is need of statistical or numerical techniques to recover the missing values in the dataset.