For the path planning problems of an unmanned ground vehicle (UGV) in complex unknown environments, a combined path planning algorithm based on improved D* Lite and jump point search (JPS) algorithms is proposed in this paper. Firstly, the jump rule and forced neighbor filtering rule of the JPS algorithm are modified, and obstacle and terrain constraints are integrated into the cost functions of D* Lite and the JPS algorithm. A threshold combination strategy based on the UGV detection range is developed in order to incorporate the two algorithms. Finally, the generated paths are optimized by removing redundant paths using jump point information and smoothing the paths with third-order B-spline interpolation. Several different environments are constructed in the Gazebo simulation software. The experimental results show that the proposed algorithm can plan secure UGV paths in complex environments with faster planning speeds and lower memory consumption, and is suitable for path planning in dynamic unknown environments.