Abstract: To minimize non-machining traversal movements in porous large-scale components, this paper proposes a numerical control (NC) forming trajectory sequentialization method based on heuristic scheme evolution (HSE). Firstly, contours and fillable regions in conceptual design scheme planning domain are generated for each NC forming surface, leading to the construction and classification of contour forming objects and surface forming objects. By employing a nearest neighborhood search approach, the initial connection sequence and starting points for different forming objects within a forming surface are determined. Subsequently, a global evolution of the connection sequence between similar types of forming objects is performed using a depth-first heuristic computation method. This involves establishing an orienteering hybrid graph (OHG) and solving it to obtain new forming object connection sequences, which are efficiently stored in the form of a static linked list. The problem reconstruction and solution refinement of initial positions and connection sequences for objects of the same type could be enhanced through both individual local evolution and multi-depth local evolution, employing a streamlined local heuristic sampling approach. Finally, the trajectories between different types of forming objects are connected to obtain the sequenced NC servo forming scheme. The results from the case study indicate that, in forming the TPMS manifold, the proposed method achieves a traversal movements optimization efficiency of 16.07%, with a maximum optimization effect of 35.83% for a single forming layer. Comparative performance testing of the HSE method against other approaches is conducted using a mechanical disk-cover-type porous component, and physical experiments are employed for validation, demonstrating that this method can save manufacturing time and improve processing efficiency.