You have to hand it to the Zelda series. It found a formula that works and exploited it to its fullest. Even the wacky Majora's Mask still used the same base that was established with the first Zelda game. It's understandable, then, that people find Zelda 2 so jarring after being conditioned to think of Zelda as stuck within the confines of this format. There is an overhead view, but it serves the same function as a JRPG overworld map. The action itself takes place in several side-scrolling setpieces with a greater emphasis on combat and a complete lack of puzzles. And the newly-implemented experience system added an extra layer of character progression, completing the RPG makeover. The style was quickly abandoned, but it remains a vastly underrated game that is begging to be revisited in a sequel. So instead of waiting for a revival that will never come, let's try and come up with a worthy sequel right now.
Aside from a graphical overhaul in the vein of Odin Sphere, the basic premise for the sequel would be the same as in the original Zelda 2. Link would be tasked with clearing seven palaces and conquering the final boss within the Grand Palace at the end. The structure itself would be essentially unchanged, with an RPG-like overworld meshing with side-scrolling action stages. After all, these elements are what made Zelda 2 so special in the first place. But the sequel would need to have a bit more polish. As great as Zelda 2 is, it suffers from unclear goals and some rough level design. Before we can realistically make a worthy sequel, we have to deal with these flaws.
The overworld was a great way of displaying the world at large and sped up traversal immensely. But for all its apparent openness, intended progression is essentially linear. However, you can tell there was an intended element of openness in the way the map is built. In order to remedy this, we will take a cue from Dragon Quest and scatter the palaces across an open map where they can be tackled in almost any order the player chooses. However, the actual flow of the game would be more akin to Castlevania: Symphony of the Night. Most of the world is open to you from the beginning, but you still need additional abilities to open up additional areas.
Progression in the side-scrolling levels would also depend on your abilities, but there wouldn't be too many branches for much of these segments. Much like the original game, the non-linearity would come from the map segments. Of course, the biggest example of this is the Death Mountain quest for the Hammer, but there were few other examples of this. The sequel would indeed have more Death Mountain-like areas to explore as well as a few branches in the stages themselves for good measure. The palaces would be the exception to this, as they will be as open and twisty as a miniature Dracula's Castle, complete with platforming and items.
In order to increase level complexity, there should be a greater focus on platforming, including tightening up the mechanics to better suit the increased emphasis on jumping. The down-stab would play a greater role than ever before, mimicking the pogo move in Ducktales to provide greater mobility. The stages themselves would also include several puzzles that would put the game a bit more in line with the Zelda series at large, though the puzzles would feel more at home in something like Braid thanks to the side-scrolling perspective. While Zelda 2 treated dungeon items as merely glorified keys, the sequel will actually see Link use them as both additional ways to maneuver through the levels and ways to solve puzzles. The items themselves would include the hookshot, the bow, and a time-altering device similar to one or more of the levels of Braid.
Of course, combat would still play a huge role in the game. One of the greatest strengths of Zelda 2 was the increased complexity to sword fighting, augmented by several combat-oriented magic spells. This same dynamic would still play a big role in the sequel, but additional enemy types would be added to create more varied combat situations. One enemy would be a bouncing knight that could be dispatched by attacking it with an upward stab, with all the other new enemies gaining similar mobility enhancements that the player would have to deal with. Also, bosses would be a compromise between the pure combat of Zelda 2 and the puzzle-like bosses of the rest of the series. Players will need to make smart use of their tools, but the predictable “three hits to kill” rule won't apply.
The core of Zelda 2 is still solid, so drastic changes to the base game aren't necessary. It would just be a matter of polishing the rough edges and adding a lot of new mechanics to the established formula. It's unlikely that we'll ever see this dream game, but we can take solace in the fact that the games Zelda 2 influenced could form the basis for a worthy follow-up.