With the recent development of machine learning, there are emerging attempts to study literature with computers. In this study, we intend to analyze the SF story dataset using a computational method. Using topic modeling, an unsupervised learning technique, we investigated the pattern of word use in science fiction stories. The produced topics could be interpreted as illustrating the subject materials in the novel and how they are used. In addition, based on the resulting topic distribution, we proposed metrics that allow us to evaluate authors’ topic use. The main contribution of this study is to present an alternative to literature study through computational methods. With advancements in computational linguistics, this study can be further expanded to illuminate other aspects of literary studies.