Abstract: 【Purpose/significance】In the era of big data, digital hoarding behavior is becoming more and more serious, especially among graduate students in school. The increasing digital hoarding behavior has seriously hindered the improvement of the scientific research ability of the graduate students in school, and this paper guides the graduate students in school to accumulate data in an or⁃ derly and effective manner through the identification of digital hoarding behaviors, so as to improve the data literacy of the graduate students in school and promote their faster adaptation to the academic research atmosphere.【Method/process】In view of disorderly ac⁃ cumulation of data, low utilization rate, and insufficient data transformation ability of graduate students to be improved, this paper col⁃ lects 56 pieces of data through interview method, uses Grounded Theory to identify digital hoarding behavior of graduate students in school, and constructs a recognition model of digital hoarding behavior of graduate students in school based on the S-O-R model.【Re⁃ sult/conclusion】Finally, 16 categories and 56 concepts were obtained, and a recognition model of digital hoarding behavior of graduate students in school was formed, the influencing factors and performance of digital hoarding behavior of graduate students in school were clarified, and their digital hoarding behavior was classified, and some targeted suggestions were given to graduate students, such as raising awareness of shared data、conducting regular psychological interviews、and improving data accumulation habits.【Innovation/ limitation】This paper applies the S-O-R model to the study of the digital hoarding behavior of graduate students in school, combine psychological concepts with information analysis methods. The next step will be to conduct targeted and extended research on the digi⁃ tal hoarding behavior of different groups in different environments, especially the in-depth quantitative research on digital hoarding behavior.