Scientific exhibition in Taiwan has been held by National Taiwan Science Education Center for more than 55 years, more than 15,000 works has been accumulated. It also becomes a famous competition for K-12 students. In order to provide recommendation of research topic to the participating students, domain experts require the growth topic and related industry from the past works. However, with the long history of the scientific exhibition, the domain experts cannot interpret all of the works in the short term. Therefore, through the computer to explore and summarize a large number of works becomes an emerging technology. In this study, we applied text-mining technology, and designed an expert rules as a computer-enabled methodology to explore the correlation between scientific exhibition and industry in the earth science and physics category. In the result, you can observe the computer programming and real estate development industries were the most growing research topic in scientific exhibition.