With　the　rapid　development　of　make-to-order　pattern　including　E-commerce　and　takeout　and　catering　service　in　restaurants,　the　study　of　integrated　scheduling　and　distribution　receives　more　and　more　attentions.　Based　on　a　practical　order　picking　and　distribution　system,　a　three-stage　hybrid　flow　shop　scheduling　problem　with　distribution　is　studied.　Each　order　is　processed　on　the　hybrid　flow　shop　which　consists　of　identical　parallel　machines　with　sequence-dependent　setup　times　at　stage　1,　identical　parallel　machines　at　stage　2　and　dedicated　machines　at　stage　3,　followed　by　a　multi-trip　traveling　salesman　problem　with　capacitated　vehicles　for　customers　of　different　destination　areas.　A　mixed-integer　linear　programming　model　is　formulated　to　minimize　the　maximum　delivery　completion　time.　A　variable　neighborhood　search　(VNS)-based　method,　a　four-layered　constructive　heuristic　method　(denoted　by　CHVNS)　and　a　hybrid　heuristic　method　(denoted　by　CONSVNS)　which　combines　the　VNS　method　and　the　CHVNS　method　are　developed　to　solve　the　problems　with　practical　size.　Computational　experiments　show　the　effectiveness　and　efficiency　of　the　proposed　methods.