The work-to-rest ratio during cycling-based high-intensity interval training (HIT) could be important in regulating physiological and performance adaptations. We sought to determine the effectiveness of cycling-based HIT with different work-to-rest ratios for long-distance running. Thirty-two long-distance runners (age: 39 ± 8 years; sex: 14 men, 18 women; average weekly running training volume: 25 miles) underwent baseline testing (3-km time-trial, V[Combining Dot Above]O2peak and time to exhaustion, and Wingate test) before a 2-week matched-work cycling HIT of 6 × 10-second sprints with different rest periods (30 seconds [R30], 80 seconds [R80], 120 seconds [R120], or control). Three-kilometer time trial was significantly improved in the R30 group only (3.1 ± 4.0%, p = 0.04), whereas time to exhaustion was significantly increased in the 2 groups with a lower work-to-rest ratio (R30 group 6.4 ± 6.3%, p = 0.003 vs. R80 group 4.4 ± 2.7%, p = 0.03 vs. R120 group 1.9 ± 5.0%, p = 0.2). However, improvements in average power production were significantly greater with a higher work-to-rest ratio (R30 group 0.3 ± 4.1%, p = 0.8 vs. R80 group 4.6 ± 4.2%, p = 0.03 vs. R120 group 5.3 ± 5.9%, p = 0.02), whereas peak power significantly increased only in the R80 group (8.5 ± 8.2%, p = 0.04) but not in the R30 group (4.3 ± 6.1%, p = 0.3) or in the R120 group (7.1 ± 7.9%, p = 0.09). Therefore, cycling-based HIT is an effective way to improve running performance, and the type and magnitude of adaptation is dependent on the work-to-rest ratio.