<p>In recent years, the exponential growth of the tourism industry has led to a surge in the number of people traveling. Efficient and accurate recommendations of tourist attractions and routes are becoming increasingly important for people’s travel. To enhance the accuracy of scenic spot recommendations, this study first designs a tourism recommendation model based on graph attention networks. Then, a clustering algorithm-based tourism scenic spot route recommendation method for scenic spot route recommendations is proposed. The data validated that the AUC values of the model based on the graph attention network were 0.8623, 0.8172, and 0.7936 respectively on different datasets, with good accuracy, recall, and F1 values. The clustering algorithm-based method had higher coverage and better diversity, with coverage rates of 0.74 and 0.48. Research has shown that tourist attraction recommendation methods based on graph attention networks can accurately recommend attractions based on user interests. Meanwhile, the clustering algorithm-based tourist attraction route recommendation method can provide users with better-personalized route recommendation services. The proposed method for recommending tourist attractions and routes effectively assists users in traveling and helps them efficiently formulate travel plans.</p>