This research aggregating the entire network of scientific research reporting information provides the users with the most suitable scientific research notification. The most important part of the scientific research report information recommendation system is the user's information and scientific research notification information, then after obtaining the two parts of information, we need to set up their own information configuration file. The key step of modeling is to extract the content from the unstructured Chinese natural language about the user information and the relevant requirements of the scientific research notification, and convert the extracted content into structured data and store it in the information configuration file. The improved text recommendation algorithm is more suitable for the recommendation of scientific research report information. The established information profile is input into the recommendation algorithm at the same time, and the similarity calculation is sequentially performed according to the user information profile and the similarity is sorted from high to low. Thus the function of scientific research report information recommendation is completed.