With the rapid development of tourism economy and the growing demand for personalized travel, the existing tourism route recommendation system has problems such as low efficiency in path planning and insensitive response to user budget and time constraints in dealing with diverse user preferences and spatiotemporal distribution of resources. To this end, this paper introduces the K-Means clustering algorithm in artificial intelligence and the genetic algorithm (GA), an evolutionary optimization algorithm, to build an intelligent route optimization system for economic tourism, in order to improve the efficiency of path generation and the ability to control details of user satisfaction. The system uses K-Means to perform preliminary clustering of the location and attribute data of scenic spots to form geographic clusters. It then uses the clustering results as a heuristic initial solution and utilizes genetic algorithms to achieve multi-objective path search optimization within and outside the cluster, taking into account multiple constraint dimensions such as time windows, ticket budgets, and interest preferences. In terms of specific detail control, the system can dynamically adjust the upper limit of daily play time, control the cross-cluster transfer distance, balance the heat score and interest tag matching, and use the cluster center as the path switching node to reduce the overall path complexity. The experimental results show that the K-Means + GA method proposed in this study performs best in terms of path quality, with a total shortest path distance of 38.1 km and a key node overlap rate of 100%. It achieves the best results in multiple evaluation dimensions, verifying the effectiveness of the integration of artificial intelligence and intelligent perception systems in the optimization scenario of economic tourism routes.