The　sharp　development　of　maritime　activities　fosters　maritime　Internet-of-Things　(M-IoT)　applications.　However,　the　computation　resources　of　vessels　are　limited,　which　cannot　adequately　satisfy　high-quality　services　in　M-IoT.　With　this　consideration,　space-Air-ground-sea　integrated　networks　(SAGSINs)　are　introduced　in　this　work.　Besides,　a　maritime　distributed　computation　offloading　framework　is　designed.　Furthermore,　an　energy　consumption　minimization　problem　is　formulated　to　jointly　optimize　the　offloading　strategy　and　the　bandwidth　allocation.　For　solution,　a　double-iteration　penalty-limited　simulated　annealing-based　particle　swarm　optimization　algorithm　is　proposed.　Simulations　are　conducted　to　validate　our　analysis　and　corroborate　the　effectiveness　of　the　proposed　scheme.　©　1997-2012　IEEE.