It should work alright, but it looks like overkill: generally, electromagnets content themselves with minimally clean, regulated or accurate supplies, except when they are part of NMR machines, precision weighting scales, etc.
You'll want to still have at least 5mA in the Zener, worst case, at the low end of V1 and the high end of Vz.
Electromagnets will have a rather large inductance.
What are you trying to stabilise? If it is field strength, then a constant-current source will likely be rather better...
What are you trying to stabilise? If it is field strength, then a constant-current source will likely be rather better...