The voltage-gated proton channel (Hv1) plays an essential role in numerous biological processes, but a detailed molecular understanding of its function is lacking. The lack of reliable structures for the open and resting states is a major handicap. Several models have been built based on homologous voltage sensors and the structure of a chimera between the mouse homologue and a phosphatase voltage sensor, but their validity is uncertain. In addition, differing views exist regarding the mode of proton translocation, the role of specific residues, and the mechanism of pH effects on voltage gating. Here we use classical proton hopping simulations under a voltage biasing force to evaluate some of the proposed structural models and explore the mechanism of proton conduction. Paradoxically, some models proposed for the closed state allow for proton permeation more easily than models for the open state. An open state model with a D112-R211 salt bridge (R3D) allows proton transport more easily than models with a D112-R208 salt bridge (R2D). However, its permeation rate seems too high, considering experimental conductances. In all cases, the proton permeates through a water wire, bypassing the salt-bridged D112 rather than being shuttled by D112. Attempts to protonate D112 are rejected due to its strong interaction with an arginine. Consistent with proton selectivity, no Na+ permeation was observed in the R2D models. As a negative control, simulations with the Kv1.2-Kv2.1 paddle-chimera voltage sensor, which is not expected to conduct protons, did not show proton permeation under the same conditions. Hydrogen bond connectivity graphs show a constriction at D112, but cannot discriminate between open and closed states.