With rapid expansion of the Internet and increasing amount of time users spend online, the Internet evolves from entertainment environment towards highly dynamic and flexible business medium. Online advertisement has become one of the most successful business model for Internet environment. There are two major types of online advertisement: sponsored search and contextual display advertisement. This paper dedicated on contextual display advertisement. Generally, contextual advertisement implementations based on topical or keyword-based relevance approach. This study addresses the mechanism of advanced contextual advertisement based on opinion about specific topic within content of webpage. Use of Natural Language Processing and Sentiment Analysis aims to determine the writer's attitude towards particular topic as: positive, negative, or neutral. This approach helps to develop an advertisement system that is more content-sensitive and consequently has higher ROI of marketing.