Abstract: The gathering system in a gas field is a multi-level network structure and its design is usually in the mode of hierarchical optimization. In the process of its calculation, the topographic relief is ignored and the obstacles are usually simplified into convex polygons, so the calculation result is more different from the practical engineering. In order to optimize the design scheme, the overall optimization model of multi-level star-shaped gathering pipeline network was established with the shortest total length of the pipeline network as the objective function and the affiliation between the stations of different levels, the processing capacity and the spatial position limitation as the constraints. In this model, three-dimensional terrain and obstacle factors are introduced, including a large number of discrete variables and continuous variables, and the constraints are nonlinear. In order to improve the global optimization ability, the group intelligence technology which combines ant colony algorithm and particle swarm algorithm together is used to obtain the solution, so as to provide an efficient and accurate method for the topology optimization of pipeline network in complicated terrains. The example calculation results show that three-dimensional terrain and obstacles affect the pipeline orientation, station position and pipeline network length. The actual surface characteristics can be simulated and the optimal obstacle avoidance path can be planned on the three-dimensional camber. The length of the gathering pipeline network based on three-dimensional terrain and obstacles is increased by about 20%.