Taking kingledion's suggestion of an alignment of the two moons with the special constellation, I have some suggestions.
For the alignment of moons to happen at least once a year and at the same date(s) each year, both of the moons have to orbit an exact or whole number of times during a year. Thus a moon could orbit 10.000 or 11.000 times per year, not 10.49 times per year, for example. It would be a vary rare coincidence for both the moons to orbit whole numbers of times per year, but there are so many possible rare coincidence that rare coincidences do happen.
It has been calculated that a moon needs to orbit more than 9 time's during a year of it's planet, in order to orbit close enough to the planet to have a stable orbit for astronomical and geological time frames. Thus the outermost moon must orbit at least 9.000 times per year and the inner moon must orbit at least 10.000 times per year.
If the inner moon orbits faster and more times per year than the outer moon, it should catch up with the outer moon at least once per year. In order to make the two moons lining up in the specified direction a unique yearly event, the lining up the two moons should happen as few times as possible.
One way to make that unique would be to have the orbits of the two moons highly titled relative to each other. The two orbital planes would intersect in a line through the planet with intersecting nodes 180 degrees apart. If the two moons line up and pass each when they are not at the nodes, they might pass at ten degrees, 45 degrees, possibly even 90 degrees, so their passing each other would not be very noticeable. But if the two moons pass through the nodes at the same time, they will pass very close and the inner moon might eclipse the outer moon.
So James may want to have the orbit of the outer moon be highly inclined, perhaps because it is a captured celestial object, and with one of the nodes where the orbital planes of the two moons intersect pointed at the desired constellation.
I think that I have discovered a couple of methods to make the alignment of the two moons happen only once per year.
1) make the inner moon orbit only one more time per year than the outer moon.
If the outer moon must orbit at least 9.00 times per year, make it orbit 9.00 times per year and the inner moon orbit 10.00 times per year. Then the outer moon will travel 3240 degrees in one year to come back to zero (or 360) degrees and the inner moon will travel 3600 degrees in one year to come back to zero (or 360) degrees. 10 orbits is 5 X 2 orbits. 9 orbits is 3 X 3 orbits. Since none of the factors of 9 is identical with a factor of 10, the two moons should not line up in the original direction more than once in a year.
What if the outer moon orbits 11 times a year and the inner moon 12 times a year? 11 x 12 is 132. In one year the outer moon will orbit 11 times and travel 3960 degrees to wind up at 0/360 degrees while the inner moon will orbit 12 times and travel 4320 degrees to wind up at 0/360 degrees. Thus they will only line up in their original direction once per year.
If the outer moon orbits 12 times per year and the inner moon orbits 13 times per year, the outer moon will travel 4320 degrees in 12 orbits and the inner moon will travel 4680 degrees in 13 orbits to both windup at 0/360 degrees.
2) make the inner moon orbit a prime number of times per year. A prime number cannot be evenly divided except by one and itself. All prime numbers are odd numbers, but not all odd numbers are prime numbers. The lowest prime numbers are 3, 5, 7, 11, 13, 17, 19, 23, 29, 31, 37, 41, 43, 47, 53, 59, 61, 67, 71.
The orbits of some bodies have orbital resonance (which sometimes destabilize and sometimes stabilize orbits) if their number of orbits during a time period have a simple numerical relationship.
As near as I can tell none of the orbital relationships that allow for a once a year alignment of moons would be resonant, though some could be close to a resonant orbit.
James will have to decide if he wants the two moons to align pointed at the proper constellation during the day or the night. During the day people would not be able to see the stars of the constellation, though ancient astronomers and astrologers could calculate the position of the sun among the invisible stars in daylight over three thousand years.
This disadvantage would be offset by the possibility of the two moons eclipsing the sun, or transiting over it, if one of the nodes where their orbits intersected was pointed at the sun at the time when the two moons passed through the node.
Depending on the apparent diameters of the two moons, they might eclipse the sun or merely transit it as tiny black dots against its brightness. Thus watchers on the ground might see the sun blotted out in two directions as the two moons crossed it in different directions, or two tiny dots crossing it in different directions. The two moons could cross the son simultaneously or minutes apart.
If the two moons passed close to the sun but didn't cross it they would be invisible in the glare of the sun. If they crossed each other in the day sky far enough from the sun, they would appear as two thin crescents. The farther from the sun, the fatter the two crescents would be.
On the other hand, if the event happens in the night sky, the two moons would look like fatter crescents, or half moons, or gibbous moons, or full moons. People on the ground could see the stars of the culturally important constellation, except in so far as the light of the two moons might drown the stars out. Possibly there is a nearby open star cluster in the important constellation that has many stars that appear far brighter than Sirius does on Earth.
I think that James should prefer to have the two moons cross paths either in the day eclipsing the sun or at night opposite the sun as full moons.
Thus it seems like it should be fairly easy for James to have someone calculate the orbits to make everything work out.
.