In　the　stockyard　of　dry　bulk　materials　such　as　coal　and　iron　ore,　multiple　stacker-reclaimers　operate　collaboratively　to　enhance　stockyard　productivity.　However,　effectively　planning　the　routes　for　these　interconnected　machines　presents　a　significant　challenge.　We　propose　a　new　modeling　framework　for　this　conflict-free　routing　problem　in　which　the　reclaiming　process　is　modeled　in　a　so-called　time-space　network　(TSN)　framework.　The　formulated　optimization　problem　is　mixed　integer　programming　(MIP),　which　has　been　proven　NP-hard.　To　address　its　computational　efficiency,　we　develop　a　two-level　metaheuristic　algorithm　to　simplify　the　encoding　complexity　of　the　original　problem.　In　the　developed　two-level　algorithm,　a　task　priority　ordering　candidate　is　listed　at　the　top　level,　while　the　detailed　conflict-free　routes　are　constructed　at　the　bottom　level　using　an　insertion-based　conflict-free　reclaiming　and　customized　route　improvement.　The　developed　metaheuristic　is　tested　on　a　large　number　of　instances　in　comparison　with　the　Gurobi　solver　and　three　commonly　used　methods　for　such　a　routing　problem.　The　experimental　results　show　that　the　proposed　algorithm　outperforms　the　other　four　methods　regarding　the　solution　quality　and　the　computation　time.　©　2004-2012　IEEE.