With the development of intelligent tourism systems, traditional tourism recommendation and guided tour services can no longer meet the increasingly personalized needs of users. Existing tourism recommendation systems face the challenges of data processing and personalized push, and cannot adapt to the dynamic needs of users in real time. To this end, this paper introduces a method that combines augmented reality (AR) technology with clustering algorithms, aiming to provide more accurate personalized tourism services by analyzing user behavior data and interest preferences. By classifying user interests through the K-means clustering algorithm, the system can adjust and display AR content that matches user interests in real time, thereby improving the experience and service quality of virtual guided tours. By refining user behavior data, interest group division and AR content display. The experiment is compared by setting up a recommendation system that uses a combination of AR technology and clustering algorithms and a traditional navigation system based on popular attraction recommendations. The data shows that the number of clicks in the experimental group ranged from 20 to 35 times, with an average of 28.3 times, while the number of clicks in the control group ranged from 10 to 18 times, with an average of 14.8 times. This shows that the interaction frequency of users in the experimental group is significantly higher and their participation is significantly stronger.