The vestibular inner ear transmits head-motion information to the brain via two populations of primary vestibular ganglion neurons (VGNs) which differ in the regularity of action potential (AP) timing (regular and irregular) and represent rate and temporal encoding, respectively. Although voltage-gated sodium (NaV) currents drive the rising phase of APs, their contributions to AP waveform and spike timing regularity in VGNs are not fully understood. NaV currents through a given α subunit can have different modes: they can be transient (inactivating), persistent (noninactivating), and/or resurgent (following relief from inactivation block). Here we consider how these NaV current modes influence AP firing in isolated VGNs, which respond to current injections with sustained and transient firing patterns. These correspond to regular and irregular spiking, respectively. Whole-cell recordings from mouse VGNs revealed that while all had large transient NaV currents that were blocked by 1 μM tetrodotoxin (TTX), some also had persistent current, and a subset had resurgent current. A blocker of NaV1.6 channels (4,9-anhydro-tetrodotoxin, 4,9-ah-TTX) partially blocked transient and eliminated persistent and resurgent currents, indicating that a substantial portion of each flow through NaV1.6 channels. In current clamp, 4,9-ah-TTX decreased excitability: increasing current threshold for spiking in all VGNs, but hyperpolarized resting membrane potential in sustained (regular) VGNs. A sodium channel agonist (Anemonia viridis toxin 2) had the opposite effect of increasing excitability as measured by decreased current threshold for spiking and increased AP rate and spiking regularity. We did not detect a clear relationship between NaV conductance and firing patterns due to unexpected variability in NaV conductance in VGN. We determined that sustained-A firing VGNs (which respond to current injections with tonic spike trains) have higher amplitude APs and faster rates of depolarization relative to transient VGNs (which fire a single spike at stimulus onset). Using the action potential clamp technique, we observed that sustained VGNs have less NaV current flow during the depolarization phase of the AP, due to their increased excitability and decreased input resistance relative to transient VGNs and possible differences in underlying channel kinetics. However, it is difficult to experimentally distinguish the effects of these currents on spiking because we lack methods to isolate current modes. We therefore modified a computational conductance-based model of VGN by adjusting and adding expressions for transient, persistent, and resurgent NaV components with properties based on our data. In model transient (irregular) VGN, adding persistent and resurgent current components had a negligible effect on firing. In model sustained (regular) VGN, adding resurgent currents reduced refractory periods and increased spike rate. When examining the impact on regularity, we observed that increasing transient current increased spiking regularity independently of increased spike rate in all model VGNs, while persistent and resurgent currents did not enhance spiking regularity. In summary, transient NaV current increases excitability and spike timing regularity by counterbalancing inexcitability driven by increased low voltage-activated potassium. Resurgent NaV current facilitates excitability in regularly firing VGNs, but do not impact regularity.