The
GURPS Space (4th) system generation sequence will work fine if you plug some other masses and luminosities into the later steps rather than using the procedures in steps 16, 17, and 18. I use actual astronomical data for real stars (from the
XHIP catalogue), and the only difficulty I have is that you need a figure for the system age at a few places (e.g. the Ocean➞Garden transformation of the world type in Step 25; volcanic and tectonic activity in Step 31). And then for realism, the age and mass of the star determine its luminosity class and luminosity, so you can't just generate a figure at random.
One thing that people often fail to mention is that the ice line and life zone of a star and the black-body temperatures of its planets are not determined by its visual luminosity (the figure cited as unqualified "luminosity" in astronomical usage) but by its bolometric luminosity. You ought to use bolometric luminosity, not visual luminosity, throughout the
GURPS Space (or any other) star system generation sequence. Use visual luminosity only for calculating the illumination of planet surfaces.
Much depends on whether you have players who know a bit about astronomy and who check your figures, or who use your figures to work out further values. I was scarred for life because my gaming group in the late Eighties included an astrophysics student, and a few other science students who were fans of "real" SF. Otherwise it is as Anders says. The science only matters to the extent that anyone is going to check your figures.