The purpose of this study is to find the main phrases of interest topics in old research by using text mining centering on domestic journals published in the national DB for 23 years. 1327 papers were used in the analysis, and the journal was extracted from the social science research including the word ’aging, old’ in the abstract, and the topic modeling method was used to show trends during that period. The most important aspect of topic modeling is to broaden and diversify the focus of aging research. As a result, five topics were identified, and the research trends and themes were confirmed through each topic. In the past, research on aged research has been limited to a micro-approach, and this study examines aged research with a more macroscopic eye. The methodology is also based on the Latent Dirichlet Allocation (LDA) approach Using analytical methods, we advanced the analysis one step further and focused on core word research that we have not been able to examine in the meantime. Through this, we were able to know what percentage of the abstracts were covered by the abstracts, and the popularity and trends of the topics could also be predicted if we understood the proportion of themes. The most important goal of this analysis is to maintain important noun phrases while removing common but unfavorable noun phrases such as ‘research’ or ‘analysis’, and to calculate the term frequency-inverse document frequency (TF-IDF) of extracted noun phrases, It shows the percentage of frequency after showing the most used words in a specific abstract. As a result, a total of 25 major noun phrases could be created for each of the five topics. Through these results, we will present the trend of aged research and future direction of research.