In recent years, urban areas have become increasingly vulnerable to inland flooding, driven by extreme weather events and aging drainage infrastructure. Unlike riverine flooding, inland flooding often occurs suddenly and unpredictably, posing serious risks to both pedestrians and drivers. Therefore, comprehensive investigations that address both riverine and inland flooding are essential. Our laboratory has previously examined the effects of flood flow velocity and inundation depth on evacuation decision-making during flood events. However, these studies have predominantly focused on pedestrian scenarios. The present study extends the analysis to driver scenarios and conducts a fundamental examination of risk perception regarding inundation depth while operating a vehicle. Specifically, a virtual reality (VR) environment representing an urban area is employed to simulate flood events, enabling an investigation of the extent to which the depth of floodwater observed from inside the vehicle differs from its actual depth as perceived by participants. As a result, it was revealed that participants perceived the inundation depth as deeper than its actual measurement.