Feature selection is an important part of machine learning and data analysis, as it helps to reduce redundancy, remove irrelevant data, and improve learning accuracy. It also reduces the computational cost, improves the accuracy of the classifier, reduces dimensionality and predicts the model. The field of Multi-Criteria Decision Making in feature selection has seen significant advancements in recent years. MCDM feature selection techniques are utilized in many ways based on various considerations like preferences, tradeoffs, weights etc. However, the efficiency and performance of feature selection algorithms remain challenges. This paper address comprehensive understanding of the motivation, methods, and evaluation criteria of feature selection is needed. And also addresses the merits and demerits of various advanced features selection techniques such as MCDM. This paper is concluded by summarizing the how much work is carried on the MCDM during past 10 years and also highlighting the future scope.