Heart rate (HR) monitoring, which reflects exercise intensity and environmental factors, is often used as the basis for pacing strategies in a marathon race. However, it is difficult to obtain appropriate feedback for only the HR value since cardiovascular drift (CV drift) occurs during prolonged exercise. Recently, cardiac cost (CC, which is HR divided by running velocity) has been shown to be a potential index for evaluating CV drift during a marathon race. The aim of this study was to clarify the relationship between CV drift and performance in a marathon race. Fourteen male university student runners participated. Each took part in incremental tests on a treadmill and subsequently ran a marathon. CV drift was evaluated using differences between CC in the 0-5 km segment of the race and every 5 km segment (ΔCC) thereafter. Marathon performance was examined from two viewpoints: absolute performance (average running velocity during the race, V_mar), and relative performance (V_mar against velocity corresponding to the ventilatory threshold, vVT achv). Significant correlations were found between ΔCC and vVT achv in the 25-30 km, 30-35 km and 35-40 km race segments (r = − 0.672, − 0.671 and − 0.661, respectively), suggesting that excessive CV drift had a negative impact on relative performance. Based on our results, we conclude that suppression of CV drift after 25 km is an important factor for improving relative performance in a marathon race.