With the increasing number of digital data and new technologies, business trends and model are becoming more dependent on computer science for making business decision and operation. By extracting user opinion for decision making and market analysis sentiment analysis is one of fast growing and reliable tool nowadays. Having a number of good research, the major issues of sentiment analysis have been solved except sentiment detection from the noun. Sometimes noun has sentiment base on the application, in this research, we have solved this with a possible solution for duel sentimental words. With existing technology, we applied some rules and developed a model to do this.