The development of the information age provides rich behavior and text data online, and it brings great opportunities for psychological research. Music is an important component of human daily life, and the current number of online music users hits 576 million (CNNIC, 2019). Previous study affirmed that personality was closely related to internet behavior and music preference. Recently, Qiu et al. (2019) found that the user's personality could predict words in their favorite music. Based on the research, the present study obtained real data of participants from their homepage of NetEase Cloud Music to increase the efficiency and ecological validity. The purpose of the present study was to explore the correlation between personality and internet behavior; the relationship between personality and lyrics characteristics on both general and specific lexical levels. In the present study, 568 participants completed the Chinese Big Five Personality Inventory brief version (CBF-PI-B) through online and offline recruitment. After obtaining the user's informed consent, the network crawler technology was used to collect data of all participant from Netease Cloud Music. Finally, 380 participants were retained after filtering to analyze the correlation between personality and internet behavior, and those who were inactive or closing down their personal homepage on the platform are screened out. Among the 380 participants, 266 participants who had 8 Chinese songs at least on personal music charts were used to analyze the relationship between personality and lyrics preference. The average number of Chinese songs is 15.07(SD=3.89). Firstly, the internet behavior data was analyzed. It showed that the personality traits of users were weakly related to their internet behavior on the music platform. The correlation value was between -.10 to .17. There was no significant correlation between neuroticism and any kind of internet behavior. Conscientiousness was found to be negatively correlated to “isSelf-introduction”, a characteristic that describes whether to present self-introduction or not, r=-.10, p<.05. Also, conscientiousness was correlated to the number of pure music in top 20, r=.10, p<.05. Agreeableness was positively correlated to the number of following, r=.10, p<.05. Openness had a positive correlation with the number of collected song lists and the average length of name of the collected song lists, r=.17, p<.01, r=.13, p<.05. Extraversion was positively related to the length of nickname, r=.15, p<.01. Then, LIWC and keyword extraction were used to analyze the lyrics data. On the general level, 13 of 43 LIWC variables were significantly related to at least one dimension of the big five personality. The correlation value was between -.19 to .13. Openness showed the most correlation with LIWC variables, while there were no significant correlation between agreeableness, extraversion and any LIWC variables. On the specific level, some keywords were consistent with personality characteristics to some extent, such as “memories” in neuroticism (r=-.13, p<.05), “years” in conscientiousness (r=.15, p<.05), “leave” in agreeableness (r=-.21, p<.01), and “forever” in openness(r=-.16, p<.01). In conclusion, the internet behavior of online music participants as well as LIWC characteristics of their favorite lyrics were related to their personality, and the keyword features of lyrics data was consistent to their personality characteristics to a certain extent. The study belongs to the intersection of psychology and information science, the results may contribute to music personalized recommendation system, music therapy, and health-care professionals.