I've studied mathematics for a long time and in my experience the second form (i.e. without the "es" at the beginning) is used more often or almost exclusively. So I would use the second version although the first one seems perfectly correct and understandable.
EDIT: Since there seems to be some disagreement here, I picked 4 random books about mathematics and computer science from my bookshelf:
- "Sei G eine endliche Gruppe von Automorphismen eines Körpers L, und sei K ihr Fixkörper." (M. Artin, "Algebra", Birkhäuser Verlag, 1998, S. 633)
- "Sei v ein Multiplikatorsystem vom Gewicht r/2 bezüglich einer Kongruenzgruppe \Theta" (E. Freitag / R. Busam, "Funktionentheorie", Springer, 2000, S. 363)
- "Sei G ein Graph mit den Cliquen C_1, ... , C_m" (C. Beierle, "Methoden Wissensbasierter Systeme", Vieweg, 2003, S. 341)
- "Es sei K ein Körper, p(x) ein separables Polynom in K und E ein Zerfällungskörper für p(x)." (E. Artin, "Galoissche Theorie", Verlag Harry Deutsch, 1988, S. 67)
So, only in one out of four books the "es" at the beginning was used. Although this selection is by no means representative, it suggests a certain trend, which I also observed in all the mathematics classes i had to take in the past.