In complex terrains, Delay-Tolerant Networks (DTN) often face challenges such as unstable node connectivity and frequent path disruptions, where efficient path planning for mobile nodes is essential to ensuring network performance. Traditional methods suffer from high computational overhead and limited adaptability, while typical swarm intelligence algorithms tend to converge to local optima, reducing routing efficiency. To overcome these limitations, this paper proposes a hybrid approach combining the Cuckoo Search (CS) and Grey Wolf Optimizer (GWO) algorithms to enhance global search capability and algorithmic stability. Curvature pruning and cubic B-spline techniques are further employed for path optimization, improving trajectory smoothness and environmental adaptability. The proposed method effectively reduces path length and collision frequency, as confirmed by simulations in both 2D and 3D environments. This improvement is crucial for enhancing DTN routing stability and overall performance. The approach is particularly suitable for efficient path planning in resource-constrained scenarios.