This study proposes a methodology to determine the optimal cross-sectional design of a high-occupancy vehicle (HOV) facility for safety performance by evaluating the effects of each relevant design element on safety and tradeoffs among multiple design variables. This aspect of geometric design is critical because it is often difficult to acquire sufficient right-of-way for retrofitting HOV lanes to existing freeway systems with recommended cross-sectional design. Detailed collision data of concurrent-flow buffer-separated HOV lanes, along with their geometric features and traffic flow data, were analyzed to establish collision predictive models by injury types for HOV and the adjacent general-purpose lanes. These models were used to determine the set of cross-sectional design elements that minimizes the expected collision occurrences. As a case study, a real freeway corridor where a conversion from contiguous to buffer-separated types was recently completed was selected to demonstrate the applicability of the proposed method. The evaluation through this case study shows that the proposed methodology is useful in determining the cross-sectional design of HOV facilities for safety based on selective use of available geometric space.