As the world grows in complexity and huge amount of data generates time to time, data analysis become difficult. In recent years, large amount of data are collected for research purposes. Such data set consists of hundreds or thousands of features. Many of the features in such data are useful information relevant to the problem. It also contains irrelevant information. So to extract relevant information a pre processing step called Feature Selection is used. Feature selection techniques like wrapper, filter, and embedded techniques are used. In Feature Selection process the relevant data are filtered to reduce the complexity before applying data mining techniques. Data mining is the process of discovering hidden, previously unknown and useful patterns essential for solving problems. For discovering classes of unknown a data mining technique called Classification is used. There are different method for classification like Bayesian, decision trees, rule based, neural networks etc. This paper analysis some existing and popular feature selection algorithms and classification.