U.S. railroad accident rates have declined substantially since the 1980s; however, further improvement in train safety remains an important objective of the railroad industry. In this paper, we describe a framework developed to assess the cost-effectiveness of railroad infrastructure improvement to reduce railroad train accidents. Higher FRA track classes have been shown to be statistically correlated with lower accident rates, thereby indicating potential safety benefits. However, such infrastructure improvement also increases both capital and operating costs for track maintenance. We use accident data from the U.S. DOT Federal Railroad Administration (FRA) accident database and cost data from several recent U.S. railroad infrastructure maintenance projects presented in an FRA report to quantitatively evaluate the safety benefits and costs associated with infrastructure improvement decisions. Our model is intended to consider the trade-off between reduced accident rates and increased costs in evaluating railroad risk reduction strategies and operational decisions. The benefit-cost analysis framework is illustrated by considering the upgrade of track class 3 to class 4 in a hypothetical case study.