To provide tourists with more personalized scenic spot route planning scheme, this study combines multi-dimensional information and A* algorithm to construct an improved intelligent scenic spot route planning model. The model first introduces multi-dimensional environmental semantic information such as terrain undulation, tourist density, emergency events, and popularity of scenic spots, and then constructs an improved dynamic road network data model. Then, the A* algorithm is improved using the lightning search algorithm to improve the dynamic path planning capability. Empirical analysis is conducted on the intelligent planning model for tourist attraction routes constructed. The planned scenic spot route had 5 turning points, with a total length of 0.9 km. It was smoother than that of other models and better meets the planning requirements for the shortest path. When the iteration was 15, it entered the convergence stage and the execution time was 2.8 s, which was significantly lower than the execution time of other models. The tourist satisfaction score was 9.4 points, and the expert satisfaction score was 8.9 points. It was higher than the satisfaction scores of other models and better meets the actual needs. Based on the above results, the intelligent planning model for tourist attractions based on heuristic search algorithms has better performance than other comparative methods, providing tourists with a more personalized travel experience and promoting the tourism industry towards a more intelligent and sustainable direction.