This study focuses on thoroughly examining how the geometric characteristics of rural roads in Peru impact safety, using the Highway Safety Manual (HSM) as the primary reference point. By adjusting the predictive HSM model to fit specific conditions, the most significant variables in anticipating road incidents were identified. Out of the 14 considered variables, three emerge as the most relevant in accident prediction. The presence of rumble strips, the Hazard Index in sections with obstacles and barriers account for 47.74% of relevance, while variables related to horizontal curves, such as length and radius, contribute with 18.35% importance in this predictive calculation. This study emphasizes the need to expand the database with information from other roads sharing similar characteristics. This would not only improve the accuracy of the calculation but also confirm the priority of th...