Abstract: Objectives: With the gradual opening of low-altitude airspace and the rapid advancement of electric vertical takeoff and landing technologies, low-altitude tourism has developed into a promising new form of travel. However, route planning for low-altitude sightseeing in urban environments faces multiple difficulties: The difficulty in quantifying landscape appreciation values, dense building clusters, fragmented feasible flight zones caused by no-fly or restricted-flight regulations over special functional areas, and the unstable performance of traditional metaheuristic algorithms in such complex scenarios. A low-altitude tourism route planning algorithm based on an improved centered collision optimizer is proposed to obtain a multi-objective balanced optimal path in complex environments. Methods： For low-altitude tourism scenarios, an objective function is established incorporating four key dimensions: Landscape appreciation value, noise impact, flight cost, and flight smoothness. Aiming at the drawbacks of the original centered collision optimizer (CCO) (including uneven initial population distribution, crude boundary handling, and insufficient local search capability), an improved CCO (ICCO) is proposed with four major improvement strategies: Optimizing the initial population using Sobol sequence and opposition-based learning, employing a Levy flight position update strategy to enhance global search capability, a boundary correction mechanism to improve convergence stability, and a differential evolution strategy to strengthen local fine-tuning ability. For large-scale tourist airspace, a block-wise planning followed by smooth connection strategy is adopted: The ICCO algorithm is used to compute the optimal path for each sub-region, and the segmented paths are further integrated into a complete flight route using cubic B-spline smooth connection techniques. Results: In comparative experiments conducted across six urban sub-regions (A—F) with varying terrain complexity and constraint density, ICCO consistently found the optimal solution fastest and outperformed genetic algorithm, particle swarm optimization, grey wolf optimizer, and the original CCO in convergence speed, solution accuracy, and stability. The standard deviation of ICCO was the smallest across all environments, and in environment C it reached zero, indicating perfect repeatability. In environment E, which featured the most complex terrain and demanded the highest multi-objective balance, ICCO achieved an optimal fitness value of –1 098.20, which was 26.77% lower than GA, 8.30% lower than GWO, 47.72% lower than PSO, and 0.24% lower than CCO. Ablation experiments confirmed that each of the four improvement strategies contributed positively to the overall performance. A preliminary assembly of the routes generated by ICCO revealed angle violations at the connection nodes (climb angles or turn angles). After smoothing these nodes using B-spline interpolation, a complete low-altitude tourism route satisfying all angle constraints was obtained. Conclusions: The combination of improvement strategies enhances the algorithm’s performance. The proposed low-altitude tourism route planning method based on the ICCO algorithm can effectively generate smooth, economical, and visually appealing low-altitude tourism routes. It possesses high practical value and can provide reliable technical support for the practical implementation of low-altitude tourism.