In this thesis we have address of fake sentiment analysis and it is a very important solution of problem that a lot of people face it because there are a lot of people write fake comments and those action might interfere accuracy of sentiment results and reviews charts so we have found technically a solution for this problem especially when fake comment and fake news have the same negative effect on society and the fast way of propagation that mislead the users and create a trust crisis , emotion and opinion manipulation. This model has been developed for one complex purpose: “Fake Sentiment analysis” The model have 3 Phases and the third phase combined these first two phases in order to achieve the solution that we worked on it in this thesis: Fake comment detection then Sentiment analysis then Fake sentiment analysis Fake comment detection phase has achieved accuracy score 0.93 and pre cision 0.94 also recall 0.94 and f1 0.94 Sentiment analysis phase has achieved accuracy score 0.95 and precision 0.96 also recall 0.96 and f1 0.96 Fake senti ment analysis phase has achieved accuracy score 0.95 and precision 0.95 also recall 0.95 and f1 0.95 Our results of the fake sentiment analysis model delivered a solution that could be enhanced in future in addition of what we have been achieved in this thesis.Here we are not doing a typical sentiment analysis but we are doing high quality sentiment analysis. Technically we are checking the sentiment by 3 sentiment analyzer not single one like the other models.