Currently, people pay more and more attention to their own health condition, and have shifted from the traditional passive participation in sports to the active participation, which brings the national sports boom is more and more intense. In this paper, we believe that the landscape design of sports scenes should take into account the needs of the public's life and the development trend of changing from landscape form to focus on the harmonious relationship between the human body and the environment, and try to put forward the form of sports landscape in the field of landscape design of sports scenes, and combined with the personalized feedback technology to design an ideal sports landscape design program with a high degree of perception to do a preliminary investigation. This requires designers to provide personalized feedback services, so this paper introduces user interest subclasses, analyses the user's historical behavioral operations, obtains the user's rating value on the interest subclasses, and then calculates the user's interest distribution based on the rating value. Based on the idea of user interest subclasses, considering the local similarity of users, we spread the traditional collaborative filtering algorithm to different interest subclasses for the prediction of overall similarity calculation and rating. Then we combine the calculation results with the user's interest distribution to obtain the final recommendation list of the user. Based on the personalized feedback technology in the sports scene landscape design provides a new perspective and way of thinking, and we hope that the sports scene landscape can bring people a pleasant sports experience, but also can call on more people to strengthen the importance of sports and fitness.