In the wireless sensor field, it is key to improve user experience and resource utilization to study personalized recommendations based on a knowledge atlas for different nodes. However, due to the limitations of wireless networks in wireless networks, existing network recommendation methods are difficult to apply. Under this background, according to the characteristics and constraints of perceptual nodes in mobile Internet, a personalized recommendation method is designed based on collaborative filtering. On this basis, through real-time monitoring and analysis of all kinds of sensor information in the sensor network, and its use habits and preferences of users, so as to achieve personalized recommendations. The results show that the cooperative filtering method proposed in this paper can be applied to wireless networks with perceptual capability. The prediction accuracy of this algorithm can reach 89.19% when recommending to different numbers of students, which improves the efficiency of algorithm execution and user satisfaction. Through the above research, the application of cooperative filtering in wireless sensing systems is proven to be feasible and effective. In future research, the method will be improved to improve the real-time and accuracy, so as to better meet the increasing demands of the intelligent library.