The lunar rover is a multi-function, mobile robot equipped with a mission. Under real terrain driving conditions, in addition to selecting the optimal path from the start point to the target point, the robot should take into account the terrain, obstacles, and other influencing factors. The main influencing factors of the terrain are steep slope gradients and slope orientation; other factors are classified as slip. These greatly increase the length and time complexity of path planning as well as the overall safety of the robot. The traditional ant colony algorithm seeks the optimal solution in path planning, but it also encounters problems such as slow convergence speed, high time complexity, and unbalanced optimization. It does not consider factors such as slip and terrain when applied to lunar rover path prediction. It is easy to fall into a local optimal solution when dealing with path planning problems. This paper proposes an improved ant colony algorithm for path planning based on the slope gradient and slope orientation for 3D raster terrain. By applying a consistent pheromone heuristic factor and pheromone volatilization coefficient, changing the terrain parameters for slip prediction, and obtaining a comprehensive cost function based on slip prediction, the traditional ant colony algorithm is improved. The influence of the comprehensive cost function based on slip prediction on the path length, convergence speed, time complexity, and iteration number of the improved ant colony algorithm is analyzed. Finally, experimental simulation data is used to verify that the improved ant colony algorithm is more effective in addressing slip prediction path planning problems.