In 2003 Michael Robinson (in The Storms of Providence: Navigating the Waters of Calvinism, Arminianism, and Open Theism) put forward the following model for how God’s knowledge of our future free choices is not incompatible with the libertarian nature of those choices. He mentions that this model is essentially argued for by both Brian Leftow and David Hunt.

This is verbatim from his book (pp. 229-232). See whatcha think! (The bold title is what Robinson calls it.)

--------------------------

The Conditional Will-Actual Foreknowledge Model

…let us suppose that humans are able to make genuinely free choices (at least in many cases) but also that (1) God knows all possibilities, (2) God formulates potential divine reactions to each of these possibilities, (3) through timeless knowledge (or perhaps temporal foreknowledge) God directly knows which possibilities in fact actualize throughout time, and (4) in light of his knowledge of (2) and (3), God also decides and knows which responses he makes throughout all time.

Propositions (1) and (2) above suggest that God has an exhaustive and meticulous plan for every possible world that he might create. Proposition (1) implies that God might know something like this:

(5) If I create W1 at T1, then at T2 events A, B, or C will follow.

(6) If I create W2 at T1, then at T2, events D, E, or F will follow.

In turn, for each of the possible events listen in the consequences of (5) and (6) God also will know the possible events that could flow from them. And so, for example, God would know that

(7) If A at T2, then at T3 events l, m, or n will occur.

(8) If B at T2, then at T3 events x, y, or z will occur.

…In turn, proposition (2) implies that for each possibility, God could envision a potential divine response. And so, God might will conditional decrees something like this:

(9) If A at T2, then at T3 I will do a.

(10) If B at T2, then at T3 I will do b.

God’s reactions in (9) and (10) may be as unintrusive as permitting the whole set of possibilities that flow from A or B, or they may compromise a more restrictive decree that permits only one of the possibilities that could flow from A or B. For example, in the former case God might decree that

(9a) If A at T2, then at T3 either l, m, or n will be permitted.

In the latter case, God might decree that

(9b) If A at T2, then at T3 only l (but not m or n) will be permitted.

…such a scheme would be quite general in the sense that through it God would not know how in fact he would accomplish his goals, only that eventually through one course or another he would achieve them.

It is here that the significance of proposition (3) emerges. For when combined with claims (1) and (2), proposition (3) implies that God knows not only how he could accomplish the divine goals, but also how he in fact does (will) achieve them. In knowing what possibilities in fact actualize for each temporal moment, God knows which divine reactions to perform at each subsequent temporal moment and, consequently, knows how the divine plan is in fact accomplished. For example, if God eternally wills that

(9) If A at T2, then at T3 I will do a.

(10) If B at T2, then at T3 I will do b.

And if God eternally knows that

(11) A at T2.

Then God also knows that

(12) At T3 I will do a.