Although many countries are revising vehicle inspection policies to improve road safety, existing standards seldom incorporate quantitative evidence on mechanical degradation at the component level. This study develops a mileage-based probabilistic failure probability model for vehicle headlamps using more than 99 million national inspection records collected in South Korea between 2014 and 2022. Headlamp luminous intensity is selected as a representative safety-critical indicator because it is legally regulated, directly measurable during inspection, and closely associated with nighttime traffic safety. The analysis indicates that the probability of headlamp failure increases sharply with vehicle age and cumulative mileage, with more than 70% of high-mileage vehicles estimated to have a high probability of not meeting the legal inspection threshold. This pattern exposes a critical limitation of the current inspection regime, which applies uniform inspection intervals without sufficiently reflecting actual usage intensity or component life characteristics. Using the estimated failure probabilities, the study examines how inspection schedules, maintenance recommendations, and targeted oversight of high-failure-probability vehicles can be restructured on a probability-oriented basis. Furthermore, a data-driven monitoring framework is introduced to identify vehicle groups with recurrent inspection failures, providing a practical instrument for transport regulators and inspection service providers. By integrating large-scale inspection data with interpretable failure probability modeling, this study establishes an empirical foundation for more efficient allocation of inspection resources, strengthened preventive maintenance, and the formulation of vehicle inspection policies and governance guided by failure probability assessments in the road transport sector. • Analyzed 99 M national inspection records (2014–2022) for headlamp failure probability. • Developed quadratic regression model for mileage-based brightness degradation. • Estimated headlamp failure probability using one-tailed normal distribution. • High-mileage vehicles (>200,000 km) show >70% threshold failure probability. • Supports probability-based inspection policies and business management strategies.
Min et al. (Mon,) studied this question.