The　X-architecture　Steiner　Minimum　Tree　(XSMT)　is　the　best　connection　model　for　multi-terminal　nets　in　global　routing　algorithms　under　non-Manhattan　structures,　and　it　is　an　NP-hard　problem.　And　the　successful　application　of　Particle　Swarm　Optimization　(PSO)　technique　in　this　field　also　reflects　its　extraordinary　optimization　ability.　Therefore,　based　on　Social　Learning　Particle　Swarm　Optimization　(SLPSO),　this　paper　proposes　an　XSMT　construction　algorithm　(called　SLPSO-XSMT)　that　can　effectively　balance　exploration　and　exploitation　capabilities.　In　order　to　expand　the　learning　range　of　particles,　a　novel　SLPSO　approach　based　on　the　learning　mechanism　of　example　pool　is　proposed,　which　is　conductive　to　break　through　local　extrema.　Then　the　proposed　mutation　operator　is　integrated　into　the　inertia　component　of　SLPSO　to　enhance　the　exploration　ability　of　the　algorithm.　At　the　same　time,　in　order　to　maintain　the　exploitation　ability,　the　proposed　crossover　operator　is　integrated　into　the　individual　cognition　and　social　cognition　of　SLPSO.　Experimental　results　show　that　compared　with　other　Steiner　tree　construction　algorithms,　the　proposed　SLPSO-XSMT　algorithm　has　better　wirelength　optimization　capability　and　superior　stability.　©　Springer　Nature　Switzerland　AG　2020.