This research paper investigates the secrecy performance in a vehicular network. A legitimate vehicular source aims to transmit confidential information to a legitimate vehicular destination, while a passive vehicular eavesdropper attempts to intercept this transmission. To enhance the practical relevance of the system model, the probabilistic existence of an obstructing vehicle is considered. This obstructing vehicle can potentially disrupt communication links between the source and the destination, the source and the eavesdropper, or both. To mitigate the adverse effects of obstruction and eavesdropping, a reconfigurable intelligent surface (RIS) is employed. Additionally, the source employs a power-splitting scheme to generate artificial noise (AN), a physical-layer security technique to degrade the wiretap channel. In this power-splitting scheme, the source’s transmission power is divided into two portions: confidential transmission and generating the AN signal. Closed-form expressions are derived for the outage probability, upper bound intercept probability, and upper bound secrecy outage probability (SOP), along with their asymptotic expressions. The influence of power-splitting ratio, transmission power, number of reflecting elements of the RIS, density of the obstructing vehicle, and the channel conditions on the performance metrics is investigated through numerical results. The results are extended to encompass practical impairments, specifically examining the impacts of node mobility, channel state information feedback delay, and AN leakage at the destination. It has been observed that an increase in the density of large obstructing vehicles leads to a reduction in the SOP. However, this degradation can be mitigated by increasing the number of passive reflectors of RIS. Additionally, based on the closed-form upper bound expression of the SOP, a power allocation optimization problem is solved to minimize the SOP in terms of the power-splitting ratio. The accuracy of the derived closed-form expressions has been verified through Monte Carlo simulations.