The aim of this study was to investigate the metabolic responses to variable versus constant-intensity (CI) during 20-km cycling on subsequent 5-km running performance. Ten triathletes, not only completed one incremental cycling test to determine maximal oxygen uptake and maximal aerobic power (MAP), but also three various cycle-run (C–R) combinations conducted in outdoor conditions. During the C–R sessions, subjects performed first a 20-km cycle-time trial with a freely chosen intensity (FCI, ∼80% MAP) followed by a 5-km run performance. Subsequently, triathletes were required to perform in a random order, two C–R sessions including either a CI, corresponding to the mean power of FCI ride, or a variable-intensity (VI) during cycling with power changes ranging from 68 to 92% MAP, followed immediately by a 5-km run. Metabolic responses and performances were measured during the C–R sessions. Running performance was significantly improved after CI ride (1118 ± 72 s) compared to those after FCI ride (1134 ± 64 s) or VI ride (1168 ± 73 s) despite similar metabolic responses and performances reported during the three cycling bouts. Moreover, metabolic variables were not significantly different between the run sessions in our triathletes. Given the lack of significant differences in metabolic responses between the C–R sessions, the improvement in running time after FCI and CI rides compared to VI ride suggests that other mechanisms, such as changes in neuromuscular activity of peripheral skeletal muscle or muscle fatigue, probably contribute to the influence of power output variation on subsequent running performance.