With　the　rapid　development　of　Industry　4.0,　the　industrial　cyber-physical　systems　(ICPS)　are　expected　to　realize　the　digital　sensing,　automatic　control,　and　refined　management　in　smart　factories.　However,　limited　bandwidth　resources　and　severe　industrial　interference　make　it　difficult　to　meet　the　real-time　and　ultrahigh　reliability　in　edge　computing　(EC)-based　next-generation　industrial　automation　networks.　To　tackle　these　challenges,　in　this　article,　we　propose　a　real-time　transmission　optimization　scheme　to　accelerate　EC.　First,　we　establish　a　hierarchical　system　model　for　smart　manufacturing　and　automation　scenarios.　Then　we　present　a　power　control　optimization　method　based　on　noncooperative　game　to　alleviate　interference　and　reduce　energy　consumption.　Finally,　we　propose　a　path　optimization　scheme　based　on　Q-learning　for　low-latency　and　ultrahigh　reliability　transmission　requirements.　Extensive　simulation　results　reveal　that　our　proposals　perform　better　in　terms　of　transmission　delay　and　packet-loss　rate　compared　with　traditional　methods,　and　therefore,　contributes　to　EC　deployment　in　ICPS.