In order to understand the nature of the lenses that generate microlen
sing events, one would Like to measure their mass, distance, and veloc
ity. Unfortunately, current microlensing experiments measure only one
parameter of the events, the characteristic timescale, which is a comb
ination of the underlying physical parameters. Other methods are requi
red to extract additional information. Parallax measurements using a s
atellite in an Earth-like orbit yield the projected velocity of the le
ns: <(nu)over tilde> = nu/(1 - z), where nu is the transverse velocity
(speed and direction) of the lens relative to the Earth-source line o
f sight, and z is the ratio of the distances to the lens and the sourc
e. A measurement of <(nu)over tilde> could distinguish between lenses
belonging to the bulge and disk populations. We show that for photomet
ric precisions of 1%-2%, it is possible to measure the projected speed
, <(nu)over tilde>, to an accuracy of less than or equal to 10% for ov
er 70% of disk lenses and over 60% of bulge lenses. For measuring the
projected velocity <(nu)over tilde>, the percentages are 40% and 30%,
respectively. We find lines of sight greater than 2 degrees away from
the ecliptic are preferable, and an Earth-satellite separation in the
range 0.7 AU-1.9 AU is optimal. The requirements of the satellite for
measuring the projected velocities of events toward the bulge are simi
lar to those for measurements toward the LMC.