Ubiquitous presence of internet, advent of web 2.0 has made social media tools like blogs, Facebook, Twitter very popular and effective. People interact with each other, share their ideas, opinions, interests and personal information. These user comments are used for finding the sentiments and also add financial, commercial and social values. However, due to the enormous amount of user generated data, it is an expensive process to analyze the data manually. Increase in activity of opinion mining and sentiment analysis, challenges are getting added every day. There is a need for automated analysis techniques to extract sentiments and opinions conveyed in the user-comments. Sentiment analysis, also known as opinion mining is the computational study of sentiments and opinions conveyed in natural language for the purpose of decision making. Preprocessing data play a vital role in getting accurate sentiment analysis results. Extracting opinion target words provide fine-grained analysis on the customer reviews. The labeled data required for training a classifier is expensive and hence to over come, Domain Adaptation technique is used. In this technique, Single classifier is designed to classify homogeneous and heterogeneous input from di_erent domain. Sentiment Dictionary used to find the opinion about a word need to be consistent and a number of techniques are used to check the consistency of the dictionaries. This paper focuses on the survey of the existing methods of Sentiment analysis and Opinion mining techniques from social media.