In　this　paper,　to　tackle　the　optimal　tracking　control　problem　for　discrete-time　systems,　an　advanced　online　value　iteration　(VI)　algorithm　is　developed.　First,　the　derivation　of　traditional　VI　for　the　optimal　tracking　control　problem　is　revisited.　Second,　the　stability　condition　of　traditional　VI　is　established,　in　order　to　evaluate　the　current　iterative　tracking　control　policy.　Third,　based　on　the　concept　of　attraction　domain,　the　improved　tracking　control　policies　under　online　VI　can　be　obtained　by　judging　the　location　of　the　current　tracking　error.　Finally,　we　prove　the　stability　of　the　online　VI　algorithm　for　the　present　tracking　control　problem.　A　simulation　example　is　displayed　to　verify　the　effectiveness　of　the　developed　algorithm.