Here's a way to do it entirely online if you don't want to download the CSPICE libraries. Visit http://wgc.jpl.nasa.gov:8080/webgeocalc/#NewCalculation and click position finder:
Fill out the data like this:
Note that a planetocentric latitude of 0 means the ecliptic latitude is 0, since we're using the ECLIPDATE (ecliptic of the date) reference frame.
The results are all the times the moon crosses the ecliptic between 1970 and 2040. This includes both ascending nodes and descending nodes.
Of course, you don't want the times, you want the actual positions, so click "Save All Intervals":
Go back to the calculation menu and hit "state vector":
Fill it out as follows:
Note that you may need to drag your saved intervals list to the "List of Intervals" box, or it may automatically be filled in.
After filling out, hit calculate:
And there you have it: the list of lunar positions for when the moon is at its descending and ascending nodes.
Notice the right ascension of the node moves somewhat linearly:
but not exactly.
If you were to plot the node's ecliptic longitude, you would see a more linear fit because of the consistent cycle between the moon's draconic and sidereal months.