To　realize　renewable　and　self-sustainable　energy　supply　in　island　region,　based　on　geographical　characteristics　with　abundant　renewable　resources,　an　optimal　model　for　island　micro　energy　grid　(MEG)　is　designed　incorporating　biomass　waste　energy　conversion　system　(ECS),　desalination,　and　power-to-hydrogen　(BSP-MEG)　Firstly,　the　mathematical　model　is　designed,　including　models　of　power　generators,　ECS,　desalination　and　power-to-hydrogen　(P2H)　devices,　etc.　Next,　the　multi-objective　scheduling　optimization　model　is　designed,　containing　conventional　scheduling　model　(Scheduling　optimization　objectives　and　constraints　established　with　minimum　operation　and　environment　costs)　and　stochastic　scheduling　model　(Minimum　Conditional　Value-at-Risk　objective　specific　to　volatility　and　uncertainty　of　renewable　generations　based　on　robust　stochastic　optimization　method).　Then,　to　solve　the　multi-objective　optimization　problem　(MOP),　a　hybrid　differential　evolution　algorithm　is　proposed　based　on　local　optimal　and　external　archiving　strategies.　Finally,　the　MEG　of　YongXing　Island　is　selected　as　an　example.　The　results　show　(1)　BSP-MEG　effectively　realized　multi-energy　cooperative　optimization,　and　promote　intra-day　peak　shaving.　(2)　BSP-MEG　reduced　operating　costs,　environmental　costs　and　Conditional　Value-at-Risk　(CVaR)　by　78.2%,　61.8%　and　77.9%　respectively,　while　curtailment　rate　by　25.6　to　0.9%.　(3)　Whether　in　general　scenario　or　worst,　BSP-MEG　can　realize　self-production　and　self-sale　of　energy　and　material,　of　which　risk　resistance　ability　is　better.　(4)　By　designing　local　optimal　and　external　archiving　strategies,　hybrid　differential　evolution　algorithm　performs　better　in　solving　complex　MOP.　In　general,　the　optimization　model　proposed　in　this　paper　can　improve　the　utilization　of　renewable　resources,　alleviate　the　shortage　of　fresh　water,　and　help　realize　renewable　and　sustainable　energy　supply.　©　2023　Elsevier　Ltd