This　paper　studies　the　multi-period　mean-variance　(MV)　asset-liability　portfolio　management　problem　(MVAL),　in　which　the　portfolio　is　constructed　by　risky　assets　and　liability.　It　is　worth　mentioning　that　the　impact　of　general　correlation　is　considered,　i.e.,　the　random　returns　of　risky　assets　and　the　liability　are　not　only　statistically　correlated　to　each　other　but　also　correlated　to　themselves　in　different　time　periods.　Such　a　model　with　a　general　correlation　structure　extends　the　classical　multi-period　MVAL　models　with　assumption　of　independent　returns.　The　authors　derive　the　explicit　portfolio　policy　and　the　MV　efficient　frontier　for　this　problem.　Moreover,　a　numerical　example　is　presented　to　illustrate　the　efficiency　of　the　proposed　solution　scheme.