In addition to being the most magnetic objects in the known Universe, magnetars are the only objects observed to generate fast-radio-burst-like emissions. The formation mechanism of magnetars is still highly debated and may potentially be probed with the magnetar velocity distribution. We carried out a 3 yr long astrometric campaign on Swift J1818.0−1607, the fastest-spinning magnetar, using the Very Long Baseline Array. After applying the phase-calibrating 1D interpolation strategy, we obtained a small proper motion of 8.5 mas yr−1 mag and a parallax of 0.12 ± 0.02 mas (uncertainties at 1σ confidence throughout the Letter) for Swift J1818.0−1607. The latter is the second magnetar parallax and is among the smallest neutron star parallaxes ever determined. From the parallax, we derived the distance
9.4
−
1.6
+
2.0
kpc, which locates Swift J1818.0−1607 at the far side of the Galactic central region. Combined with the distance, the small proper motion leads to a transverse peculiar velocity
v
⊥
=
48
−
16
+
50
km s−1—a new lower limit to magnetar v
⊥. Incorporating previous v
⊥ estimates of seven other magnetars, we acquired
v
⊥
=
149
−
68
+
132
km s−1 for the sample of astrometrically studied magnetars, corresponding to the three-dimensional space velocity
∼
190
−
87
+
168
km s−1, smaller than the average level of young pulsars. Additionally, we found that the magnetar velocity sample does not follow the unimodal young pulsar velocity distribution reported by Hobbs et al. at >2σ confidence, while loosely agreeing with more recent bimodal young pulsar velocity distributions derived from relatively small samples of quality astrometric determinations.