<p>The surgical cutting guide, fabricated through fused deposition modeling (FDM), provides an efficient solution for mandibular defect reconstruction. However, it suffers from problems such as weak strength, poor surface quality, and long fabrication time. This paper aims to find the optimal setting of process parameters (viz. print speed, layer thickness, raster width, and extrusion temperature) to simultaneously improve conflicting responses in terms of surface roughness, tensile strength, and print time for FDM-printed mandibular cutting guide. Experiments were performed based on response surface method (RSM) utilizing 30 sets of PLA-printed specimens. Regression models were developed using artificial neural network (ANN), whose architecture was tuned by particle swarm optimization (PSO) to improve predictive accuracy. The three-objective problem was solved by multi-objective particle swarm optimization (MOPSO) to get the optimal Pareto set. Technique for order preference by similarity to ideal solution (TOPSIS) combined with analytical hierarchal process (AHP) was applied to find the most preferred solution. Compared with experimental results, the optimized solutions exhibited prediction errors of 3.24% for surface roughness, 3.73% for tensile strength, and 2.95% for print time, respectively. This highlights the efficiency of the proposed methodology in effectively optimizing conflicting objectives for the mandibular cutting guide, which will help promote the application of 3D-based surgical guides in the medical field.</p> Graphical abstract <p></p>