Sentiment analysis is an application of Natural Language Processing (a branch of Artificial Intelligence) that revolves around detecting the sentiment of text. A common dimension for measuring sentiment uses labels positive, negative and neutral, as well as many other possibilities (i.e. how strong the sentiment is, how active vs subdued it is, etc). This thesis illuminates the perennial conundrum of predicting sentiment from short pieces of text, predominantly engendered in social networking sites. The challenge becomes an overriding concern if the resulting use of the language that does not conform entirely to grammatical and syntactical rules and the constraints that the short text imposes are taken into careful consideration. The Word-Graph Sentiment Analysis Method is proposed to identify the sentiment of the tweets via the containing sequence of the words. In fact, the value of the application counterweights the endeavor difficulty allowing the social media users to quickly and automatically evaluate the public’s disposition towards multiple topics. As such, in recent years sentiment analysis has manifested itself as an indispensable methodology in marketing and other suchlike application domains.