Through semantic analysis of user evaluation texts of reading APPs, the potential needs and preferences of users are explored to provide optimization design directions for such products. Conduct user preference research through content analysis. First, select some user evaluations of high-scoring reading APPs, and summarize and organize the texts. Secondly, use ROST CM6 software to disassemble sentences of user evaluation text and calculate the word frequency. Then, code and sort the user’s needs, quantitatively analyze the user's preference when using reading APPs, clarify the optimization design direction, build a preference model for reading APPs, and apply and verify the model results. Through the user evaluation of reading APPs, it is possible to qualitatively analyze the user’s satisfaction with the product, and quantify the potential needs and preferences of users, so as to clarify the optimization design direction, cultivate users to develop good reading habits, improve the reading quality and finally improve users' usage efficiency and satisfaction.